Introduction

I first attended Mobile Tartu 2 years previously as an invited speaker. A big focus of my talk 2 years ago was on the importance of reproducibility and open science in transport research. Open science, including sharing code, data and results in accessible formats, increases trust in research findings, makes work more generalisable and ‘internationalisable’, makes it easier for others to build on research, and can help democratise transport planning (Lovelace et al. 2020). For Mobile Tartu 2026 I was invited to deliver a workshop for the PhD school that precedes the conference and I naturally made open science a key theme in the workshop. The result was a 2-day course on Data Science for Transport Planning delivered by me and my ITS colleague and PhD student Juan Fonseca, shown in the photo.

Tallinn

Before the event we got to spend some time exploring Tallinn. There was an outdoor festival taking place, which featured live music, the ideal way to acclimatise to the Estonian culture. Although I couldn’t understand a single word of the lyrics I could appreciate the catchy and brilliantly executed songs. It was only several days later that I found out that it was an Estonian-language progressive pop band called Lonitseera. Their music is available from their website, from YouTubeand beyond, highly recommended!

The next day, on Sunday 7th June, it was time to go to the PhD school venue. I found time to go for a quick run in the morning, and ended up at Linnahall, a largely abandoned venue constructed for the 1980 Olympic games. Climbing up the huge concrete structure gave me an impression of how much Tallinn, and the whole of Estonia, has changed since regaining independence following a referendum in which almost 80% of its population voted for independence from the former Soviet Union in 1991.

The Mobile Tartu PhD School

While the main Mobile Tartu conference is always hosted at the University of Tartu (founded in 1632, closely following other early universities like Oxford, Cambridge, Salamanca and Uppsala, and more than 200 years earlier than the University of Leeds where I’m based), the PhD School that precedes it moves to a new location each time. For Mobile Tartu 2026 the venue was an Ice Age museum with impressive sculptures of Woolly Mamoths and good views of the Estonian countryside on our doorstep. The venue was also on the edge of a lake (Saadjärv) which was the ideal place to cool off after taking a sauna to celebrate completing the first day’s work.

As ice-breaker activities, we were asked to create a map of the world in the space shown below and find things we had in common. This was a great way to get to know each other and to see the impressive international spread of participants.

Before the workshops began there was a talk by Andres Sevstuk, who was doing a research visit to the University of Cambridge, UK, as part of a sabbatical. (Coincidentally, as part of the research visit, Andres is organising a symposium on modelling active travel as part of the Applied Urban Modelling conference, where I will also be presenting.)

Some photos from the talk, which was excellent, are provided below (plus an embedded LinkedIn post which requires a LinkedIn account to view).

There was a good link between Andres’s talk and the topic of his workshop: modelling pedestrian movements using the Python package madina (Sevtsuk and Alhassan 2025). There were 4 workshops in total, with links to the workshop materials in the links where I could find them:

1. Data Science for Transport Planning: from origin-destination to route network datasets (Robin Lovelace and Juan Fonseca Zamora, our workshop)
   

• See the workshop homepage for links to the materials, code and slides.

2. Modelling Pedestrian Mobility in Cities: Tools for Sustainable Urban Design (Andres Sevtsuk)
   
3. Correcting Biases in Mobile Phone Mobility Data: The DEBIAS Framework (Francisco Rowe and Carmen Cabrera)
   

• This workshop was based on the in-progress debiasR R package and it’s open documentation

4. Regional Inequities in Accessibility: Comparative Analysis of GTFS and Passenger Counts (Anto Aasa, Ago Tominga and Martin Haamer)

Our workshop: Data Science for Transport Planning

Our workshop built on the Data Science for Transport Planning continuous professional development (CPD) courses we have delivered for practitioners, plus the Tools and Skills for Reproducible Transport Research that Juan and I delivered last year at the EIT Summer School in Lisbon.

This time we focussed specifically on using origin-destination data from the {spanishoddata} R package (Kotov et al. 2026). We also included demonstrations of how to use the OpenStreetMap data with R and Python to generate estimates of flow on networks. We had 4 teams of 3-4 participants and each was tasked with developing a reproducible workflow to answer a research question of their choice using the data and tools we had introduced.

A challenge for participants was to develop an understanding of the data and tools provided, while simultaneously developing their own research. It was a bit like trying to compress a research project into 2 half-days or, trying to fly a plane while building it. An additional challenge that we set the participants was to present their work using slides generated Quarto and hosted on the course website. As outlined below, every team aced it, resulting in 4 excellent presentations on the final day of the workshop, professional-looking interactive slides that others can benefit from, and the experience of committing code and opening pull requests to a GitHub repository, an impressive achievement given that many of the participants had never used GitHub, let along GitHub Codespaces or Pull Requests before.

The final presentations were as follows:

1. Leisure and Commuting Mobility Patterns in the Spain-France Border Region — Eva Hajčiarová, Paul Bairoh, Allison Fernández-Lobo, Juliette Le Corguillé. Slides
2. Cross-Border Flows Between France and Spain through Catalonia — Mariana Montero Vega, Leonardo Moreira, Juan Arquero Gallego, Keren Or Rosenbaum. Slides
3. Cross-Border Trip Purposes, Bike-Sharing & Public Transport Strikes, and OD Flow Validation — Jonas Wübbenhorst, Helen Tera, Xiao Cai, Jaanika Jaanits. Slides
4. How Is Human Mobility Driven By Amenities? A Pamplona Case Study — Zijun Ding, Thomas Dimos, Laura Altin, Emiliano Beltran. Slides

One of my favourite parts from the presentations was seeing Juan Arquero presenting the results of his team’s analysis of Spanish OD dataset to explain the travel patterns near his home in Madrid, shown in the video below.

Beyond learning about new datasets and data science tools, the workshop also taught new ways to collaborate and communicate research findings, which is a key part of the reproducible research ethos. The course repository now has 10+ contributors and should help inspire future work and teaching.

A key decision that we made early on with the teaching was to use Devcontainers to provide a consistent environment for all participants. Despite some initial hiccups, this seems to have worked well, enabling people new to coding and GitHub to get up and running quickly, and allowing us to focus on the research questions rather than technical issues. A key lesson learned for anyone thinking of using Devcontainers: setup can take time, so it’s worth using a pre-existing container and devcontainer.json settings. Use a tried-and-tested container such as the itsleeds/tds container used in this workshop, or the geocompx/docker container that it builds on, rather than trying to build your own from scratch. See the source code for our workshop for the devcontainer.json settings we used and feel free to adapt them for your own workshops.

Despite the fact that Devcontainers made installing and using the tools easier, it was still impressive to see how quickly the participants got up to speed with the tools and data, and how quickly they were able to produce high-quality outputs.

Well done to all the participants!

The Mobile Tartu Conference

The rest of the conference was great, including standard presentations by researchers from a wide range of backgrounds. The quality of the research presented was high, and there was lots to learn. There was also a poster session, providing a good way for researchers to present their work one-to-one and get feedback.

This format, with a PhD School, and then multiple ways to present research, means that Mobile Tartu has something to offer everyone.

Another great thing about the conference is how good the organising team, and the Mobility Lab in general, is at getting the work out there, meaning people who cannot attend in person can still benefit from the research presented. It’s in that spirit that I have shared this write-up, and provide some more links below.

• The Mobility Lab YouTube channel has keynotes from the conference
• The Mobile Tartu 2026 conference website has links to the conference programme and abstracts
• The madina Python package for modelling pedestrian movements, which was the focus of Andres Sevtsuk’s workshop
• Course materials for the Data Science for Transport Planning workshop delivered by me and Juan

Thanks to everyone for making it such a great event, and hope to see you at the next one in 2028!

Why migrate? Wowchemy and HugoBlox closed ecosystems

I explored options including migrating from Wowchemy to HugoBlox. If you search for Wowchemy on GitHub today, there are no public repositories under github.com/wowchemy. The project activity has moved under the HugoBlox organisation: github.com/HugoBlox.

So one path was to stay in the Hugo ecosystem and migrate templates and content there. The other path was to simplify and move to Quarto for content-first publishing and easier long-term maintenance.

I went with Quarto. Despite its cons of being slower to generate html (quarto preview can take minutes for large sites, something the team is working on in the currently-experimental quarto-dev/q2 repository) and being less flexible in some ways than Hugo, Quarto is unbeatable for scientific and reproducible content.

Converting content

For anyone sitting on years of posts, events, and publication metadata, this is the kind of utility that turns a risky migration into a manageable one.

Another useful piece of the Quarto ecosystem is the Academicons extension workflow documented here: https://rmflight.github.io/posts/2022-12-03-using-academicons-in-your-quarto-blog/. Academicons are great for academic profile links (ORCID, Google Scholar and similar), but the experience is still more extension-oriented and less integrated than the tightly packaged academic site features available in HugoBlox. That is one reason I think Quarto could benefit from more high-level, Wowchemy/HugoBlox-style extensions for academic websites.

Legacy site archive

The old version of my site remains online here:

• http://robinlovelace.net/old-site/

That gives a stable reference point while the new stack evolves.

Domain and GitHub Pages setup

With help from Jakub Nowosad, I configured my domain so GitHub Pages can serve from the apex domain setup. The practical result is that repositories with gh-pages outputs appear automatically under robinlovelace.net.

For example:

• New URL: https://robinlovelace.net/rs5c/
• Previous URL: https://robinlovelace.github.io/rs5c/

The old GitHub-hosted URL now redirects, so links continue to work while the canonical home is on robinlovelace.net.

Next steps

I’m still looking into how to better automate things. I’ve set-up GitHub Actions to create Netlify previews on each pull request, and am looking to set-up a regular workflow to run every month or so to update citations. Overall, I suggest keeping it fairly simple, with themes and extensions that are well-supported and maintained.

I’ve also found that nothing quite beats the out-of-the-box Publications page in HugoBlox for managing publications, so I’m looking into ways to automate the creation of a similar page in Quarto.

Check out the source code of my newly updated website if interested in any case and feel free to reach out, e.g. using the newly set-up Giscus comments system on this site. Hoping this new set-up will encourage me to blog more. Watch this space!

Unified schema

Older data files have columns like carriageway_hazards_historic while newer ones use carriageway_hazards. v4.0.0 detects these variants, merges them into the modern names, and drops the redundant columns.

library(stats19)
crashes = get_stats19(year = 1979:2024, type = "crashes")

Parsing fixes

read_stats19() now builds a custom parser from the CSV header, which removes the warnings about unmatched columns that appeared in previous versions. We also fixed a bug where 2024 latitude and longitude values were truncated to integers.

Missing values

Codes like -1, “Code deprecated”, and “Data missing or out of range” are now standardised to NA during formatting, so is.na() works consistently.

Performance

The package now uses readr Edition 2 by default, which supports multi-threaded parsing. Loading large files is noticeably faster.

New functions

• match_tag() joins government TAG cost estimates (RAS4001) to collision data
• clean_make(), clean_model(), and clean_make_model() standardise the 2,400+ raw strings in the vehicle dataset

Multi-year downloads

Year ranges now download bulk historic files once and filter efficiently. The 1979 file is also handled correctly (it used to be returned as a catch-all for any older year).

Feedback wanted

We plan to submit to CRAN soon. Please install, test, and report any issues:

pak::pak("ropensci/stats19")

Issues: github.com/ropensci/stats19/issues

Acknowledgements

Contributions from David Ranzolin and Adam Sparks (rOpenSci review), Malcolm Morgan, Layik Hama, and Blaise Kelly. Funding from the RAC Foundation.

Links

• GitHub
• Documentation
• Changelog

The city

The city of Copenhagen is beautiful, child-friendly and highly conducive to active travel, things that are enabled by comparatively low (but still harmful) levels of car ownership and driving in city. There are entire books on the topic that I cannot do justice to here, so I will just share some photos I took during the trip.

Overall it was an inspiring trip. There are many research ideas brewing, and I look forward to putting some of them into thought, words, code and action over the coming months. I would love to learn more about how Copenhagen and other cities with high percentages of trips made by active modes got there so feel free to get in touch on that. If you’re interested in any of the code underlying Anastassia’s thesis, check out the NERDSITU GitHub organisation.

More could be said on the research and city sides, I would love to create a photo map of my trip, but that can wait for another day.

References

Working at Active Travel England

My two-year contract at Active Travel England concluded on January 3, 2025, marking the end of a detour in my career. As outlined in a previous post, taking the job in the Civil Service and going part-time at the University of Leeds to enable it was a big decision.

The decision to go part time in both jobs was the right one though: it allowed me to follow through on the belief that publicly-funded researchers owe it to society to ‘give back’. Plus, the differences between academia and the Civil Service are not too big: both are knowledge-based institutions with a focus on evidence and the public good. (In some countries such as Germany, academics are civil servants.)

More specifically, accepting the offer to work part time in a new part of central government, devoted to a topic that I am passionate about, was exciting. It allowed me to build a data and digital team and help shape a rapidly-evolving organisation. Furthermore, I knew that the role in Active Travel England would allow me to work on projects that will have huge impacts for years to come, while continuing to work with outstanding people as part of my main career in academia.

The job at ATE didn’t come out of the blue. I started working for ATE as the Director of Data and Analysis, on an interim basis in 2022. It was a secondment from my role as a 10DS Fellow at the data science team in No 10, which was itself a secondment from the University of Leeds. Clearly, building digital and data science capacity in government was going to be easier to deliver if I was working for government directly. Positive societal impact has always been a key driver for me, and the opportunity to work on the Active Travel England project was too good to pass up.

My proudest achievements in ATE include:

• Setting-up a grant funding agreement between Active Travel England and the Alan Turing Institute (see details at turing.ac.uk and Forbes) enabling the development of new digital tools for more effective transport planning, including planning (see below) and ‘inspectorate’ tools, supplementing the spreadsheet-based Active Travel England design assistance tools.
• Instigating the ‘ATIP’ project hosted at plan.activetravelengland.gov.uk (see open source code at github.com/acteng/atip)
• The SchoolRoutes project, which builds on od2net and the Propensity to Cycle Tool to help local authorities identify and prioritise school routes for investment in active travel infrastructure.

After my experience there, I would recommend working in the Civil Service to anyone who is interested in making a positive impact on society and has the skills to do so, see job opportunities at civilservicejobs.service.gov.uk.

Academic exploits

In 2024 I continued to work on a range of academic projects, with the focus being on the Network Planning Tool for Scotland (NPT), the new BICISCHOOLS project with colleagues at the University of Lisbon, and various ongoing research projects and collaborations.

Highlights included:

• I completed two books on geographic data analysis using free and open-source programming languages for reproducible research. Both are now with CRC Press, available for pre-order, and will be published in 2025.
  • Geocomputation with R (second edition), a major update of the first edition, which was published in 2019.
  • Geocomputation with Python, a new book that builds on the success of the R book and applies the same principles to the Python programming language.
• Delivering a keynote at the Mobile Tartu conference, and participating in the summer school that preceded it, which incidentally led to a new R package [spanishoddata (see below).
• The internationalisation and extension of methods developed for the Propensity to Cycle Tool project, including:
  • Publishing our work on the Cycle Route Uptake and Scenario Estimation (CRUSE) project in an academic paper (see cruse.bike for an interactive web app resulting from the research)
  • Completion of the Biclar web application and paper, published in Computers, Environment and Urban Systems (CEUS) describing the underlying methods.
• Participating in the Spatial Data Science across Languages (SDSL) conference, which was held in-person in Prague in September 2024.

My year in open source

In 2024 I continued to contribute to open source projects, including:

• Developing a simple R interface to the od2net Rust crate, which can generate route networks for OD data in a fraction of the time of other approaches.
  • This work was part of the SchoolRoutes project, see the acteng/netgen package for more details.
• Instigating and co-authoring the new spanishoddata R package, which enables reproducible scientific research and discovery in mobility and other fields based on the world’s first open access and high-resolution national origin-destination dataset. Thanks to the Spanish government for making this data open access and to Egor Kotov, who led the development of the package after we met at the Mobile Tartu conference.
• Contributing to the parenx Python package, now on pip which provides functions for simplification of geographic representations of transport networks.
• Work on software for joining spatial networks, including improvements to the stplanr::rnet_join() function and contributions to the anime (Approximate Network Integration, Matching and Enrichment) package for fast and flexible linestring matching in R, Rust and (hopefully soon) other languages.

See my GitHub profile for more details on my open source work.

Community projects

Working across three roles as Lead Data Scientist at Active Travel England, an academic at the University of Leeds, and an author and maintainer of open-source software involved a lot of screen time. The perfect antidote to that is getting outside and doing things with other people and 2024 was pretty strong in that regard. I had three main outlets for my outdoorsy and social side in 2024: the community-run Bike Fix Sunday project to help people fix their bikes in the local community, the Chapeltown Cohousing community-led housing project where I live, and the ‘Oaty Cortado run club’.

These were a fun and rewarding outlet during downtime from work and family life and, as the photos below convey, were great fun.

Thoughts on 2025

I’m looking forward 2025, building on the foundations laid in 2024. Happy New Year everyone!

My career path

I didn’t set out to work in academia and certainly didn’t plan to become a professor. Although my career has been mostly in higher education, I think it is a little unusual in the way I’ve got there, with a big focus on impact and practical skills, rather than research outputs, administrative roles and memberships of committees and boards. As outlined in snippets from the evidence submitted to the promotion panel below, a large part of my career has been spent working on projects that have had a direct impact on policy and practice (indeed, one of the questions from the panel was whether there’s a risk that my work becomes too much like consultancy). Before going into that, here’s a brief overview of my career to date in chronological order:

• 1990-2002: Attended Weobley High School and Weobley Primary, state schools near where I grew up in Herefordshire
• 2002-2004: Attended Hereford Sixth Form College, studying A-levels in Maths, Geography, Psychology, and AS levels in Physical Education and English Language and Literature
• 2004-2008: Studied Geography at the University of Bristol’s School of Geographical Sciences, where you were allowed to do combined physical and human geography, something at the time called Environmental Geography, an option that is sadly no longer available at the University of Bristol but which is happily available in a number of other UK and international universities (source: Google Searches with the text strings "environmental geography" site:ac.uk and "environmental geography" university course)
• 2008-2009: Won a scholarship to study Environmental Science and Management at the University of York (an excellent course that I would recommend to anyone interested in the environment and sustainability, I’m glad to see that it’s still going strong according to the website)
• 2009-2013: 4 year combined masters and PhD studentship as part of the EPSRC-funded E-Futures Doctoral Training Centre at the University of Sheffield, where my supervisors Dimitris Ballas, Matt Watson, and Stephen Beck provided the skills across a range of disciplines (economics, microsimulation, policy engagement and engineering) that have been the foundation of my career
• 2013-2014: My first job, an 18 month contract working on the National Centre for Research Methods (NCRM) funded TALISMAN project on geospatial data analysis and simulation project (huge credit to Mark Birkin who gave me my first job in academia and who encouraged me to pursue my own research interests)
• 2014-2016: Research fellow in Data Analytics (University of Leeds, Leeds Institute for Data Analytics), where I got stuck into data science and started to take computer science and programming seriously, leading to the book I co-authored with Colin Gillespie, Efficient R Programming
• 2016-2019: University Academic Fellow in Transport and Big Data, at the Leeds Institute for Transport Studies (ITS), where I worked on a range of projects including the Propensity to Cycle Tool, and where I’m still based
• 2019-2024: Associate Professor of Transport Data Science at the University of Leeds (part time since 2023)
• 2023-2024: Working part time in 2 year contract for the Civil Service as Director of Data & Analysis, Head of Data and Digital and Lead Data Scientist at the new government agency Active Travel England (ATE)
• 2024 onwards: Professor of Transport Data Science at the University of Leeds

What a journey!

The promotion process and portfolio

In the UK higher education system, promotions take place through a formal process and, as in many jobs, based on a set of criteria. As outlined on the University of Leeds’ Human Resources website, there are three pathways you can take, which encourages diversity in the types of professors that are appointed (although there is still a long way to go in terms of diversity on every dimension):

• Student Education
• Research & Innovation
• Academic Leadership

As someone focussed on research and innovation, I applied for promotion through the Research & Innovation pathway, the guidance notes of which can be found here: Grade 10 Research & Innovation guidance notes. 18 criteria are listed, with 9 in each of the categories of Research and Innovation part A (RIA1 to RIA9) and Research and Innovation part B (RIB1 to RIB9). You have to provide evidence for all 9 of the RIA criteria and 3 of the RIB criteria, with a maximum word limit of 5000 words overall. You also need to submit a publication list and a CV. An important aspect of the promotion process for a professorship is endorsement from experts in the field and I am very grateful to people who know my work and who were willing to provide statements.

The panel session itself was less stressful than I expected. I had braced myself for a grilling, in between a PhD viva and a job interview, but it was more of a conversation. The people on the panel asked very good questions about my application and my future plans, and the process generated ideas that I would like to take forward, including the broad aim of “providing data science to the masses” as one of the panel members put it.

I share some of the portfolio that I’m proud of and which may be of use/interest to others below, starting with key publications and the reasons why they are important (this was a great opportunity to reflect on my favourite papers and the impact they have had, hence sharing here).

Key publications

Following in the footsteps of my colleague Kate Pangbourne, I provided a short description of each of 10 publications I could select for the submission, with the following intro (following the perhaps problematic ideas of ‘publish or perish’ and that papers are the ‘universal currency’ of academia):

I have more than 60 academic publications, including peer-reviewed papers in high-ranking journals, externally funded high-impact reports, and four books. My work is highly cited, attracting 2,450 citations (1985 since 2019), according to Google Scholar, with a strong upward trend and a H-index of 27 (25 since 2019) and an i10 index of 50 (45 since 2019). I have a strong record leading publications: I led all 10 of the publications below.

1. Lovelace, Robin, M Birkin, Joseph Talbot, and Malcolm Morgan. ‘Cycle Network Policy, Planning and Investment Transformed by the Propensity to Cycle Tool’. Research Excellence Framework. University of Leeds, 2023. https://results2021.ref.ac.uk/impact/847d1191-7f25-46ba-a399-b481125edc8f?page=1.

I led this REF Impact Case Study, from its inception and framing through to the details of which quotes to include. The case study documents the impact of work I have led on strategic cycle network planning, making millions of pounds worth of public money invested more effectively as part of the government’s Cycling and Walking Investment Strategy, part of the Infrastructure Act. It was a key component of the joint ITS and School of Geography submission and was rated as a Four-star level of impact, meaning world-leading in terms of originality, significance, and rigour (Lee Brown, leader of the submission, personal communication).

2. Lovelace, Robin, Anna Goodman, Rachel Aldred, Nikolai Berkoff, Ali Abbas, and James Woodcock. ‘The Propensity to Cycle Tool: An Open Source Online System for Sustainable Transport Planning’. Journal of Transport and Land Use 10, no. 1 (1 January 2017). https://doi.org/10.5198/jtlu.2016.862. Cited by 167

In this paper I articulate the need for a new approach to Planning Support Systems (PSS), building on previous literature, and proceed to outline a new approach based on open access origin-destination data and new web technologies allowing more participation in the transport planning process. The paper has had a major impact on transport planning internationally and has been used to support investment in better evidence for active travel planning, including in follow-on funding for projects in New Zealand, Scotland, the Republic of Ireland, and Portugal.

3. Lovelace, Robin, Jakub Nowosad, and Jannes Muenchow. Geocomputation with R. CRC Press, 2019. Cited by 204

Geocomputation with R is a seminal contribution at the intersection of the disciplines of data science, geographic information science and, in the popular Transportation chapter, transport planning. The book has been described as a “must-read if you’re into data analytics/science” in a review on Amazon (where the book is rated 5 stars with 19 reviews), “Mandatory reading for my Geoinformatics for Planning course at School of Planning and Architecture, Delhi” by Balamurugan Soundararaj, and “comprehensive and readable” by Professor Chris Brunsdon. The book is part of 11 university courses, including the modules Reproducible and Collaborative Data Science, led by Carl Boettiger, Berkeley University of California, Analysing spatial data, led by Professor Roger Bivand, Norwegian School of Economics, and Spatial Methods in Community Research, led by Professor Noli Brazil, University of California. The book has also formed the basis of income-generating knowledge exchange activities and teaching (notably as part of the Transport Data Science module) in the University of Leeds.

4. Lovelace, Robin, Mark Birkin, Philip Cross, and Martin Clarke. ‘From Big Noise to Big Data: Toward the Verification of Large Data Sets for Understanding Regional Retail Flows’. Geographical Analysis 48, no. 1 (1 January 2016): 59–81. https://doi.org/10.1111/gean.12081. Cited by 94

In this paper I outline the importance of model/observation inter-comparisons for modern data science applied to mobility datasets, with a focus on origin-destination data. The approach has played a key role in subsequent research. For example, Professor Martin Trépanier and colleagues, in a 2022 paper in the high-ranking journal Transportation Research Part A, cited the paper to explain their focus on dataset cross-validation when analysing large travel survey datasets.

5. Lovelace, Robin, and Richard Ellison. ‘Stplanr: A Package for Transport Planning’. The R Journal 10, no. 2 (2018): 7–23. https://doi.org/10.32614/RJ-2018-053. Cited by 55

This paper outlines the design and implementation of the stplanr software, an R package which has been downloaded more than 153k times and used to support more than a dozen papers, which may not have been possible without the package. The methods implemented in the package enabled the modelling of national transport network datasets and estimation of traffic volumes down to the street level in multiple high-impact projects. An example of the international impacts of the research is its use to support a study of TB patients and accessibility to TB treatment initiation clinics, as outlined in a post in the rOpenSci open science discussion forum and 2021 paper “Clinical, health systems and neighbourhood determinants of tuberculosis case fatality in urban Blantyre, Malawi”, published in the high-ranking (IF: 4.2) journal Epidemiology & Infection.

6. Lovelace, Robin. ‘Open Source Tools for Geographic Analysis in Transport Planning’. Journal of Geographical Systems, 16 January 2021. https://doi.org/10.1007/s10109-020-00342-2. Cited by 61

In this invited contribution I set out the ethical, intellectual and practical rationale for making transport modelling tools more open, reproducible and participatory. The paper has been cited by numerous subsequent papers, including in a recent review in the high-ranking journal Computers Environment and Urban Systems that summarises the paper’s findings, stating that “the review by Lovelace (2021) explores the current landscape of open source software and conclude that open source solutions can fulfill the professional needs of modern day transport planners” and echos these findings in the conclusions, supporting and amplifying the impact of the paper.

7. Lovelace, Robin, John Parkin, and Tom Cohen. ‘Open Access Transport Models: A Leverage Point in Sustainable Transport Planning’. Transport Policy 97 (1 October 2020): 47–54. https://doi.org/10.1016/j.tranpol.2020.06.015. Cited by 40

In this paper co-authored by John Parkin (Professor of Transport Engineering, University of West of England) and Tom Cohen (Senior Lecturer in Transport Planning and Management at the University of Westminster) I set-out to go beyond the concept of ‘open source’ in transport planning and articulate the concept of ‘open access transport models’, meaning tools that are not only open and conducive to reproducible research, but also accessible to the public. The paper has had a substantial impact, being mentioned in discussion panels coordinated by eminent transport modelling professional Tom Van Vuren MBE and forming a basis of an opinion piece in the industry publication TransportXtra.

8. Lovelace, Robin, Joseph Talbot, Malcolm Morgan, and Martin Lucas-Smith. ‘Methods to Prioritise Pop-up Active Transport Infrastructure’. Transport Findings, 8 July 2020, 13421. https://doi.org/10.32866/001c.13421. Cited by 29

This short paper demonstrates the cross-transferability and agility of the reproducible data science approach to transport planning advocated throughout my work. It does so by applying the approach to the urgent need to generate new evidence to support the sudden shift in government transport policy during the period of physical distancing measures during the most critical year of the COVID-19 pandemic response, with much of the work done while health professionals were in crisis response mode in the spring of 2020. The paper led to a DfT contract to develop a tool to support local authorities to rapidly develop new cycle lanes and low traffic neighbourhoods, as part of the first round of Emergency Active Travel Funding (EATF), which subsequently became the Active Travel Fund, rounds 2 to 4, which I have continued to work on in my Civil Service job for Active Travel England.

9. Lovelace, Robin, Martijn Tennekes, and Dustin Carlino. ‘ClockBoard: A Zoning System for Urban Analysis’. Journal of Spatial Information Science, no. 24 (20 June 2022): 63–85. https://doi.org/10.5311/JOSIS.2022.24.172. Cited by 10

This paper demonstrates my ability not only to use established methods but also to create new methods and contribute fundamental new ideas and algorithms that can be used in a wide range of fields. Although the paper was written for transport planning applications (we used it to support the high-impact ActDev project), it is relevant in the fields of Urban Analytics (contributing to research-led teaching in the Transport Data Science module, which is part of the Urban Data Science and Analytics MSc programme of which the module is a part), Development Planning and Urban Geography. Although a recent publication, it has been cited in numerous follow-on papers and was quoted at length in a paper in the journal Journal of Environmental Policy & Planning (Impact Factor 3.2), highlighting the importance and interdisciplinary cross-transferability new methods for zone generation outlined in the paper. The paper is also a good example of my ability to collaborate across national and disciplinary lines, with co-authors Dustin Carlino (an ex Google American software engineer currently working at the Alan Turing Institute) and Martijn Tennekes (a senior civil servant in the Dutch statistical agency) providing vital input into the implementation in open source software for others to benefit from the methods.

10. Lovelace, Robin, Roger Beecham, Eva Heinen, Eugeni Vidal Tortosa, Yuanxuan Yang, Chris Slade, and Antonia Roberts. ‘Is the London Cycle Hire Scheme Becoming More Inclusive? An Evaluation of the Shifting Spatial Distribution of Uptake Based on 70 Million Trips’. Transportation Research Part A: Policy and Practice 140 (1 October 2020): 1–15. https://doi.org/10.1016/j.tra.2020.07.017. Cited by 20

I include this paper because it demonstrates the power of new data science techniques to generate evidence for transport planning and address pressing policy questions. Building on previous findings and controversies, the paper sets out to explore the extent to which bike share schemes can be used by less privileged groups, with reference to a large dataset from Transport for London. The findings, made possible by new techniques which I have helped to develop, provide an important contribution to the debate and clearly support the view that investment in bike share schemes, and investment in active travel more broadly, can support social equity objectives. The paper is also a good example of my ability to bring together a wide range of people including practitioners and multi-disciplinary academic research teams to deliver high-impact research, with co-authors from the bike share industry (Smoove), academia and the third sector (the shared mobility not-for-profit Co-Mo).

Portfolio

The portfolio was long and detailed. Some of the key bits from it are below:

My input into the overall strategic direction of transport planning and modelling disciplines is exemplified by two recent articles in transport practitioner-focused magazines. I was interviewed for the Chartered Institution of Highways & Transportation (CIHT) on new methods in Transport Planning for their Transportation Professional magazine, leading to an article in their magazine. As outlined in RIA9 I was also interviewed in the practitioner-focused magazine TransportXtra in May 2024.

In terms of aligning with the University of Leeds’ strategies, I wrote:

My work has a strong ethical direction that aligns well with the University of Leeds’ strategies, especially its climate action strategy. As outlined in the Climate Plan, “we shall increasingly reorient our research and teaching away from the fossil fuel sector”. A question that is not asked (or answered) is “what will we orient towards, in place of the fossil fuel sector and associated energy intensive sectors?” My research agenda can help answer this question, with a solutions-focused approach that has tangible real-world impacts and potential for much greater future impacts, especially if I am supported to develop my team and apply for international funding bids.

Regarding ‘translational’ activity I wrote:

Most of my research has been funded by industry, governments, and transnational organisations, highlighting the high social, policy, and economic value of my work. National transport planning bodies including the Department for Transport (DfT), Transport Scotland (TS) and Transport Infrastructure Ireland (TII) have all funded tools which are now used in production by 50k+ people each year … as described in a successful REF Impact Case Study that I led for ITS in collaboration with staff in the School of Geography.

Regarding teaching, I referred to the Transport Data Science course that I coordinate, writing:

I have a sustained record of accomplishment leading outstanding research-led and innovative teaching, as illustrated by my ongoing leadership of the Transport Data Science (TRAN5340M) module from 2019 to present.

I also highlighted my international research projects and links, with reference to a recent keynote speech that I delivered at the Mobile Tartu conference where I gave a talk on Future-Proof Transport Planning (photo from that keynote shown above). International research has been a growing focus of my work, as highlighted in the quote from ITS’s head of school on my promotion:

After completing his PhD in transport and energy use, Robin joined the University of Leeds in 2013 working on new methods for adding value to large geographic datasets and then ITS in 2016 as a University Academic Fellow in Transport and Big Data. He instigated and led the development of the Propensity to Cycle Tool (publicly available at www.pct.bike), which has transformed strategic cycle network planning in England was the subject of a REF Impact Case Study. Robin’s work has led to open tools for more evidence-based decision-making overseas, including in the Republic of Ireland (with results publicly available at CRUSE.bike), the Greater Lisbon area (with results hosted on the website of the regional planning body at tmlmobilidade.pt) and in an ongoing project funded by Transport Scotland (resulting in an open access web application hosted at www.npt.scot).

Reflections on incentives, inspiration and future plans

An article in the Guardian titled “Work in an academic-professional hybrid role? Say goodbye to career progression” argues that it’s hard to progress if you undertake both professional and academic roles. Although the article is talking about taking on dual roles within the same institution, it applies to academics like me who have taken a step outside higher education. I agree with the article’s take that academic career progression is generally rather narrowly focussed, and that there is “a need in universities for a band of staff who bridge the two much larger groups”.

Unlike the author of the Guardian article, I have not found dual roles to be a barrier to progression, and this has been supported by the Research Excellence Framework (REF) process, a key feature of how UK research is evaluated. Despite issues with the REF, it clearly does incentivise people who are focussed on delivering outputs that go beyond papers and citations, with a blog post by Jon Collett at the London School of Economics showing that top case study submissions such as the one outlined above can lead to substantial rewards. REF also provides a chance for people to talk about how their work has be used in the real world during the promotion process, which is a good thing (my biggest issue with the process is that it’s so time consuming for academics submitting them and feels a bit like ‘marking your own homework’, could REF shift to a process in which an independent body not affiliated with any university plays a greater role, I wonder). I’m not sure what promotion criteria are like in other universities, but I do think that having pathways for academics interested in “translational” research, or research that has a direct impact on policy and practice, is important.

However, like many others, I have found having 2 jobs to be a challenge at times, while also being highly rewarding. It’s clearly a good thing when academics get out of their comfort zones and step into the ‘real world’, whether that’s in government, industry, or the third sector. However, it seems that there are disincentives to doing so: most jobs have a ‘baseline level’ of work that needs to be done, and if you take on a second job, you have to do two lots of baseline work.

One issue with academic reward systems, I think, is that they can tend to promote people who are confident and who ‘shout loud’ rather than the quiet, thoughtful types who might be doing the most important work. I was surprised at the extent to which it’s up to the promotion candidate (me in this case but because anyone can in theory be promoted at any time it can mean everyone) to ask about or (less honourably) even lobby for promotion. In fact, it was only the encouragement of colleagues and friends that led me to apply for promotion in the first place. So my message to people who know they are doing great work, but who are not sure if they should apply for promotion, is to go for it, you’re worth it!

This suggests that programmes like the UKRI policy fellowship scheme, which I benefited from and which led to my job in Active Travel England, are a good thing. It also suggests that employers should be flexible and should maybe even incentivise staff to take on secondments and other roles outside their main job. In many ways academia is the dream career path and, more than other careers, it can be seen as more of a vocation than a ‘normal’ job. There are of course downsides but these are hugely offset by the upsides. So I’m very happy to be moving back to being a full time academic next year and look forward to taking my work in higher education to the next level in 2025!

All careers have their pros and cons and I think academia is unique in the level of flexibility and autonomy it offers, especially when you get to the lofty heights of a professorship. Each time I have had my doubts about the academic career path, I have been inspired by others who are using their positions to make a difference. I am inspired by people like Edzer Pebesma, Megan Ryerson, John Parkin, Rachel Aldred, Anna Goodman, Antonion Paez, Lake Sagaris, Michael Szell, McEwen Khundi, Geoff Boeing, Alex Singleton, and amazing colleagues and students at the University of Leeds, including Greg Marsden, Kate Pangbourne, James Tate, and Jillian Anable. Thanks to you and everyone who using their academic positions and roles to support positive changes within and beyond academia in a variety of ways. In turn, I hope that some of my work will inspire others.

I’m also inspired by my amazing partner Katy, family and friends: home and happiness are foundations on which many good things can build.

I’m still processing what it means to have reached the top of the academic career ladder (in terms of the UK grading system at least, there is a long way to go in terms of the work!), but am committed to continuing to learn from others, sharing (including through this blog and through the amazing open source software community), and focussing on projects that are likely to have a positive impact as I go.

I do have more specific plans for high-impact projects, but will save that for another day.

Tuesday 11th June

Tuesday was the first day of the 2-day PhD School that preceded and then joined the main conference. Traditionally it takes place outside the main conference location and this year was no exception: a nature reserve just outside Tartu was the location chosen this year. There were fun ice-breakers, including a decentralised quiz during which participants were tasked with finding at least one person who met each of a dozen criteria relating to topics including people’s favourite GUI-based GIS software (QGIS of course!), … and Estonian puddings (the hardest criteria because it had to be a non-Estonian who had eaten and remembered the name of said hard-to-pronounce pudding).

Even more inclusive was how everyone introduced themselves. We were asked to stand in a circle looking in and say our name, where we’re from and how many ‘Mobile Tartus’ we had previously attended (number one for me, I was impressed to learn that there were people there who had all 9). After some refreshments and general mingling it was time for the first keynote talk: Kristjan … who had spent time on a placement with the Estonian government. His talk was fascinating to me, as someone who had made the shift from academia to the public sector. Instead of working on separate jobs — as I did when I shifted from being on secondment to Active Travel England as part of the 10DS Fellowship to taking on a second job in early 2023 — Kristjan seems to have gone all-in on the public sector job and achieved a huge amount. As far as I am aware, in Estonia (as in many other countries including Germany and perhaps all the nordic nations), being an academic means being a civil servant, that is to say: like civil servants, academics in Estonia are employed by the government.

I’m not sure about the distribution of pros and cons of this, but I imagine that it makes it easier for academics to take on government roles (has anyone done research into the extent to which government policies are evidence-based in countries in which academic staff in state universities are civil servants, compared with countries in which they are not, I wonder). The talk was fascinating, covering projects that had involved new and emerging datasets held by the Estonian government.

I already knew that Estonia was an early adopter of ‘digital government’ but didn’t realise the breadth and depth of digital government initiatives and policies it has in place (in fact ‘initiatives and policies’ don’t really do justice to some of these things, digital government seems to have become so embedded that it’s simply how things are done). Digital government in Estonia includes (or perhaps more accurately is underpinned by) X-Road, an open source system for secure, decentralised management of multiple government datasets that serves hundreds of thousands of queries every year, saving thousands of person-days worth of work. The system allows different parts of government to inter-connect, helping with the process of what the Department for Transport’s Chief Scientific Advisor calls de-siloisation. I took some notes during the talk, one of which was the following:

If you never scale, you’ll never have an impact on policy: as Kristjan said it’s not enough to put the concept out, write the paper and move on: you need a prototype.

After the talk (which had many other examples of digital government and its clear benefits), we split into four groups for the PhD workshop sessions:

• A team working on datasets representing refugee movements
• A group working on Mobility Hubs
• A team working on data from Helsinki provided by Hennriki Tenkannen, who led the session

Despite not being a PhD student and therefore not having signed-up to a session, I decided to attend Hennriki’s session, wanting to brush-up on my Python skills and to check-out some open access transport-related datasets from Finland.

The final part of day 1 was a social which involved optionally going into a sauna and (even more optionally) taking a dip in the river. Tired from all the activity of the first day we headed to Tartu where the rest of the conference would take place.

Wednesday 12th June

The second day involved a continuation of the PhD workshops, plus the opening of main Mobile Conference for all participants. There was plenty of time to work on the topics and other things: I decided to work on a related dataset and the nascent {spanishoddata} package during some of this time, thanks to an overdue meeting with Egor Kotov from the Max Planck Institute for Demographic Research and Universitat Pompeu Fabra.

Update from January 2025: the package is now published on CRAN, and the source code is hosted on GitHub at github.com/rOpenSpain/spanishoddata.

Thursday 13th June

The first full day of the conference involved keynote talks by Nico Van de Weghe from Ghent University and Anu Masso, interspersed with regular talks submitted by participants.

Nico’s talk was on the topic of “Exploring Possibilities: generative AI in human mobility research”, and can be seen in the video below:

Anu’s talk was on the topic of “Mobility Data Justice: Estonian Data Manifesto”, and can be seen in the video below:

Friday 14th June

This was the final day of the academic part of the conference and a big one for me, as it was the day of my talk. As shown in the video below, my talk was titled “Reproducible research and open tools for future-proof transport planning”. See more about the talk here and see the open source code at futureproof.robinlovelace.net.

As with the previous day, there were two paper sessions, one in the morning and one in the afternoon. The topics were

• Morning session
  • Session 4A: Accessibility to public transit and shared mobility
  • Session 4B: Daily use of urban space: understanding mobility and segregation with mobile big data
• Afternoon session
  • Session 5A: The spatial, digital, and temporal dimensions in mobility and urbanity

The panel discussion, which I had the honour of being involved in, was on the topic of “Understanding accessibility and socio-spatial disparities with mobile big data”, and highlighted the importance of open data and reproducible research to ensure that the benefits of new data sources are shared widely and equitably. It was refreshing to have the involvement of people not only in academia but also from the public sector and industry, including Kertu Vuks from the Tartu City Government, in the session.

Conclusion

Mobile Tartu 2024 was a fantastic conference, one of the most enjoyable and productive I have ever attended. It was an honour to present and I thank the organisers for the opportunity.

I look forward to the next Mobile Tartu conference in 2026, and hope to see many of the same faces there.

Research

It was a productive year for my research outputs, in which I was co-author on the following research outputs:

• Cycle network policy, planning and investment transformed by the Propensity to Cycle Tool, not an academic paper but a case study about the pathways from research to impact, as part of the UK’s Research Excellence Framework (REF). The work was assessed as being a 4* REF Impact Case Study according to the amazing team at the University of Leeds who supported me to condense 5+ years of work into a single report (Lovelace et al. 2023).

• A road segment prioritization approach for cycling infrastructure, a paper led by my PhD student Hussein Mahfouz, that explores methods for prioritising investment in active travel networks that take into account equity and cost effectiveness (Mahfouz, Lovelace, and Arcaute 2023). It took a LONG and occasionally frustrating amount of time to get through the peer review process but I’m really happy that this is now published in a top journal (JTG).

• Contraflows and cycling safety: Evidence from 22 years of data involving 508 one-way streets is a paper led by my former PhD student Caroline Tait (Tait et al. 2023). It is, to the best of our knowledge, the most detailed data-driven assessment of the safety of contraflow cycling interventions which is highly policy relevant: it challenges the unfounded assumption that contraflow cycleways should not be approved on safety grounds, especially in places like London where there are existing cycling communities and high potential.

• Where to invest in cycle parking: A portfolio management approach to spatial transport planning was the result of an Master’s dissertation project at ITS, it’s a solid method for prioritising in cycle parking that I hope to see implemented soon (Ito, Morgan, and Lovelace 2023).

• Road lighting and cycling: A review of the academic literature and policy guidelines is the first review into the research on street lighting and cycling led by my colleague Eugeni (Vidal-Tortosa and Lovelace 2023).

• Packaging code for reproducible research in the public sector is not actually published yet, but I’m very proud of this collaboration between me, Federico Botta from the University of Exeter, Arthur Turrell of the Data Science Campus, and Laura Gilbert from the data science team in No. 10 (10DS) (Botta et al., n.d.)

Second job

2023 was an exhilarating year work-wise, as I took on a new role as a civil servant, in addition to my main job as an associate professor in transport data science at the University of Leeds. The new part time role is with Active Travel England, the government agency tasked with creating a step change in walking and cycling nationwide. I had a broad remit but my main aim during the first part was to build the data and digital team, which I’m happy to say, job done. We started 2023 with only 1 person with a data and digital remit (me) and end it with a small but effective team of 7.

There were many proud moments over the last 12 months in Active Travel England, one of the greatest being the launch of the collaboration between Active Travel England and the Alan Turing Institute, the UK’s national institute for data science and AI. Building on my links with the ATI as a Turing Fellow, and my previous working with Dustin Carlino and others on the prototype ActDev tool, we secured funding for a 2 year project for the ATI to develop the Active Travel Infrastructure Platform (ATIP).

From the Active Travel England press release:

This [collaboration] will enable the development of new functionality in the Active Travel Infrastructure Platform (ATIP), which helps councils to map out proposed schemes and see the impact they could have locally.

These new tools will be paired with existing data sources such as OpenStreetMap, to create innovative solutions that will help build the evidence needed to meet national government’s objectives on active travel, including for 50% of short trips in urban areas to be made by walking, wheeling and cycling by 2030. The investment will demonstrate how new software engineering and data science techniques can support evidence-based planning and support Active Travel England’s mission.

I played a role in securing the funding, writing the grant funding agreement, submitting it (with lots of support from ATE colleagues) and getting people in the ATI and ATE onboard. The most credit goes to Dustin and everyone involved, it’s still a work in progress, but the first version of the ATIP is now live and being used by local authorities across the country to plan walking and cycling schemes.

See the open source code underlying the functionality plus links to a live web app at github.com/acteng/atip.

Like the Propensity to Cycle Tool (which I got to demo to the then Secretary of State for transport in 2017), the ATIP project is highly conducive to collaborative planning and live demos. We got a chance to show off the tool during a live demo with the government minister responsible for active travel, Jesse Norman (now Guy Opperman since November 2023).

Home at last

It was a more relaxing end to the year than 2022 which, as outlined in the previous year’s reflections, was largely spent unpacking. I’m happy to say that we’re now settled into our new home in the Chapeltown Cohousing project which provides an amazing base for our life in Leeds. It has a shared garden and the cars are parked outside, meaning plenty of space for Kit, now aged 2, to run around without worries about road traffic danger.

My wife Katy and I are also now the proud owners of an e-bike, an early Christmas present which is a low carbon and (thanks to the cycle to work scheme) cost effective way of getting around, allowing us to combine biking to work with the nursery run.

Hopes and dreams for 2024

In 2024 I plan to focus more on my academic research but will retain a role in ATE to support on data science and digital innovation. My dream is spend more time with my son Kit and on the work side to build a team developing cutting edge and internationally applicable research and tools for sustainable transport planning. Now settled into the new home and into my new roles, it’s all doable, just requires hard work, focus, and good work-life balance.

References

Why Ubuntu?

This article is about installing software on Ubuntu, not the operating system itself. However, it’s worth mentioning a few things about Ubuntu before diving into the software set-up that I use. There are many factors that should go into the decision of which operating system to use, although most people (except software developers) never even have the opportunity to make that decision as they are handed Windows on every computer they use (except when they use online services, most of which run on Linux systems set-up by developers running Linux, a notable example of companies not drinking their own champagne/dogfooding). That is not the topic of this blog post but worth stating the reasoning behind using Linux: from economic, educational, security and good vibes perspectives it’s a win, win, win, win. Include the user-friendliness and stability of Ubuntu and it’s a win x 5 situation!

If you don’t already use Linux and are curious about Ubuntu, I recommend checking out one of the great blogs on ‘dual booting’ linux on computers that currently have Windows

In fact, because it was from UK based Linux hardware specialists Entroware, it already had Ubuntu installed. The laptop in question: an Entroware Proteus 15”, which has great spec at a reasonable price. It’s also a looker as shown in the photo.

I may write a review of the Entroware Proteus at some point. For now, suffice to say that it’s one of the few off-the-shelf options for people wanting to buy a decent new laptop without having to pay for a clunky, resource-intensive operating system that suffers from various security vulnerabilities that Linux avoids.

Install key packages

After you have installed Ubuntu or bought a laptop that has it preinstalled, it’s time to get things set-up! The packages that are ‘key’ will vary from person to person. I’m confident that the list I have below will be useful to many people. If you like making maps, reproducible research, and using future-proof tools for data science, programming and geocomputation, I’m confident many will be useful. The guide below covers the following:

• GitHub’s CLI tool for command-line collaboration and code sharing
• R with access to pre-compiled packages, a powerhouse for statistical computing
• R packages for working with spatial data
• RStudio, a space-aged editor for R
• VS Code with extensions for R, Python, C/C++, Rust, Docker, and more
• Docker, which gives ultimate power and flexibility to install and run software

I’ve also created content for adding more amazing software, find those on the GitHub repo I set up to encourage contributions and to make the content more future proof:

• QGIS
• Rust
• Getting started with tools for geocomputation
• Other tools for boosting productivity and developer experience
  • Rust command-line utilities
    • ripgrep
    • fd
    • sd
    • exa
    • dust
  • CopyQ
  • AppImage Launcher
  • LogSeq
  • Zotero
    • Make Zotero available to the launcher
  • ddterm
  • Signal
  • Flameshot
  • Chrome
  • Edge
  • KmCaster
  • Syncthing
  • Discord
  • Slack

So let’s dive in!

Fire up a terminal, e.g. with Ctl-Alt-T after booting Ubuntu, then enter the following.

type -p curl >/dev/null || (sudo apt update && sudo apt install curl -y)
curl -fsSL https://cli.github.com/packages/githubcli-archive-keyring.gpg | sudo dd of=/usr/share/keyrings/githubcli-archive-keyring.gpg \
&& sudo chmod go+r /usr/share/keyrings/githubcli-archive-keyring.gpg \
&& echo "deb [arch=$(dpkg --print-architecture) signed-by=/usr/share/keyrings/githubcli-archive-keyring.gpg] https://cli.github.com/packages stable main" | sudo tee /etc/apt/sources.list.d/github-cli.list > /dev/null \
&& sudo apt update \
&& sudo apt install gh -y

# Log-in to GitHub from the command line
gh auth login
gh repo clone Robinlovelace/install-geocomp-ubuntu
cd install-geocomp-ubuntu 
less README.md

git config --global user.email "test@tilia.org.uk"
git config --global user.name "trachelium"

sudo apt update
sudo apt upgrade
sudo apt install -y --no-install-recommends software-properties-common dirmngr
# Add the keys
wget -qO- https://cloud.r-project.org/bin/linux/ubuntu/marutter_pubkey.asc | sudo tee -a /etc/apt/trusted.gpg.d/cran_ubuntu_key.asc
# add the R 4.0 repo from CRAN -- adjust 'focal' to 'groovy' or 'bionic' as needed
sudo add-apt-repository "deb https://cloud.r-project.org/bin/linux/ubuntu $(lsb_release -cs)-cran40/"

# You'll be asked to choose your geographical area
sudo apt install -y r-base r-base-core r-recommended r-base-dev
# Install PPA with pre-compiled packages
sudo add-apt-repository ppa:c2d4u.team/c2d4u4.0+

# My own selection among 5000...
sudo apt install -y r-cran-rgl r-cran-rjags r-cran-snow r-cran-ggplot2 r-cran-igraph r-cran-lme4 r-cran-rjava r-cran-devtools r-cran-roxygen2 r-cran-rjava
sudo apt install r-cran-tidyverse
sudo apt install r-cran-sf r-cran-tmap r-cran-osmextract r-cran-mapview
# System deps for cartography
sudo apt install -y libgdal-dev libproj-dev libgeos-dev libudunits2-dev libv8-dev libnode-dev libcairo2-dev libnetcdf-dev
sudo apt install -y libglu1-mesa-dev freeglut3-dev mesa-common-dev
# Extra packages for image manipulation
sudo apt install -y libmagick++-dev libjq-dev libv8-dev libprotobuf-dev protobuf-compiler libsodium-dev imagemagick libgit2-dev
# rspatial
sudo apt install r-cran-sf r-cran-terra r-cran-mapedit r-cran-tmap r-cran-mapdeck r-cran-shinyjs 
Rscript -e 'install.packages("languageserver")'

wget https://download1.rstudio.org/electron/jammy/amd64/rstudio-2023.03.0-386-amd64.deb
sudo dpkg -i rstudio*
rm -v rstudio*

rstudio

rstudioapi::openProject("~/Download/install-geocomp-ubuntu/")

git status
git commit -am 'Update instructions with info on using RStudio'

wget -qO- https://packages.microsoft.com/keys/microsoft.asc | gpg --dearmor > packages.microsoft.gpg
sudo install -o root -g root -m 644 packages.microsoft.gpg /etc/apt/trusted.gpg.d/
rm packages.microsoft.gpg
sudo add-apt-repository "deb [arch=amd64] https://packages.microsoft.com/repos/vscode stable main"
sudo apt update
sudo apt-get install code
code --install-extension ms-vscode-remote.remote-containers

code --install-extension quarto.quarto

code --install-extension reditorsupport.r
code --install-extension github.copilot

wget https://github.com/quarto-dev/quarto-cli/releases/download/v1.3.313/quarto-1.3.313-linux-amd64.deb -O /tmp/quarto.deb
sudo dpkg -i /tmp/quarto.deb

Rscript -e "remotes::install_github('quarto-dev/quarto-r')"

sudo apt-get install \
    ca-certificates \
    curl \
    gnupg \
    lsb-release

sudo mkdir -p /etc/apt/keyrings
curl -fsSL https://download.docker.com/linux/ubuntu/gpg | sudo gpg --dearmor -o /etc/apt/keyrings/docker.gpg
# Use the following command to set up the repository:
echo \
"deb [arch=$(dpkg --print-architecture) signed-by=/etc/apt/keyrings/docker.gpg] https://download.docker.com/linux/ubuntu \
$(lsb_release -cs) stable" | sudo tee /etc/apt/sources.list.d/docker.list > /dev/null

# Install Docker Engine

# Update the apt package index, and install the latest version of Docker Engine, containerd, and Docker Compose, or go to the next step to install a specific version:
sudo apt-get update
sudo apt-get install docker-ce docker-ce-cli containerd.io docker-compose-plugin

sudo groupadd docker
# Add your user to the docker group:
sudo usermod -aG docker $USER

# Log out and log back in so that your group membership is re-evaluated.
# If you’re running Linux in a virtual machine, it may be necessary to restart the virtual machine for changes to take effect.
# You can also run the following command to activate the changes to groups:
newgrp docker

# Verify that you can run docker commands without sudo.
docker run hello-world

# If you initially ran Docker CLI commands using sudo before adding your user to the docker group, you may see the following error:
# WARNING: Error loading config file: /home/user/.docker/config.json -
# stat /home/user/.docker/config.json: permission denied

# This error indicates that the permission settings for the ~/.docker/ directory are incorrect, due to having used the sudo command earlier.

# To fix this problem, either remove the ~/.docker/ directory (it’s recreated automatically, but any custom settings are lost), or change its ownership and permissions using the following commands:

sudo chown "$USER":"$USER" /home/"$USER"/.docker -R
sudo chmod g+rwx "$HOME/.docker" -R

gh repo clone geocompx/geocompy
code geocompy

quarto preview

Summary

This blog post outlined my experiences getting set-up with a new Ubuntu installation and provided reproducible instructions for people who want to follow some of my steps. An important feature of open source software is choice.# There isn’t just one way to do things. That can be frustrating and daunting: so many decisions to make!

However, the act of making decisions is important. Choosing the set-up that works for you, rather than just following others, is highly recommended. So I encourage you to NOT follow all these instructions verbatim. Do your own research. Try things out, uninstalling things that don’t work for you, and keeping things that do. I outline some of the thinking around software choice here and plan to update that after the dust has settled around my updated set-up documented in this blog post and associated repo.

Ultimately, technology is about increasing what you can do. So the key thing is what you do with the technology. A decent computer with some of the open source tools outlined above is a powerful technology indeed. There are many other tools to install now and many other tools will be developed in the future. Chances are high that at least some of these tools, not covered here, will be of use to you.

Another thing about open source communities is their emphasis on sharing and learning, which is what inspired me to document the steps I took and share them with the world. Feel free fork my repo, tell me about issues and suggest changes via Pull Requests or in the comments below.

One thing that I’m not sure of: how best to get a good Python environment up-and-running? For the Geocomputation with Python book project I have found that micromamba works great, but that’s in Docker and I’m not sure of the best way to set-up my default Python instance on Ubuntu. Any suggestions welcome.

Thanks to everyone for building the amazing software installed in the course of this blog post and to everyone who has helped me with instructions, blog posts, answers on Q&A sites and more. To see a slightly more complete list, see the source code underlying this block post here: https://github.com/Robinlovelace/install-geocomp-ubuntu

Jittering

Much of my work has involved processing origin-destination (OD) data. OD data underlies the highest impact project I have worked on, the Propensity to Cycle Tool (hosted at pct.bike) and follow-on projects^{1 23}. To re-cap, the PCT uses open access OD data representing trips between small administrative zones in England as the basis of a model that estimates cycling potential at zone, desire line, route and, most importantly, route network levels. As outlined in the figure below

Moving house

This last milestone of 2022 was in many ways the biggest. A house is more than just a place sleep: it can be a foundation, a source of strength, a place to call home. So it was a big moment for me and Katy when we got the keys to our new place, on 5th December.

Getting the keys was just the start of the moving process though. Thanks to Leeds City Council, and City Connect which is joining up Leeds’ patchy active travel network, we borrowed an electric cargo cycle, Raleigh’s Pro Electric Cargo Trike. With a 900 litre front carrier, this futuristic machine had the capacity to make our move a lot easier than if we had relied on a car alone.

I’m grateful to CityConnect and encourage anyone needing to regularly shift moderate loads, up to around 100 kg, to do some research into the revolutionary potential of e-cargos to increase the efficiency and fun of moving stuff. You can sign-up for a trial at leeds.gov.uk/evtrials.

References

Introduction: why Mastodon

Twitter has been purchased by Elon Musk, a prominent technologist and car salesperson with questionable judgement when it comes to social media. This, in combination with evidence of rapid spread of disinformation and hatred centralised platforms, has triggered a period of rapid change for Twitter, to put it mildly.

According to one user on the Twitter alternative Mastodon (the topic of this post):

A collapse in ad revenues has been followed with an immediate 50% reduction in employee headcount.

Brace, Brace!

Incidentally, for a taster, this is how the post looks on the platform (surprisingly similar to Twitter):

Even the founder of Twitter, Jack Dorsey, signalled his regret about the direction of travel of the platform that he pioneered:

Folks at Twitter past and present are strong and resilient. They will always find a way no matter how difficult the moment. I realize many are angry with me. I own the responsibility for why everyone is in this situation: I grew the company size too quickly. I apologize for that.

— @jack November 5, 2022

These statements and associated events create two broad reasons to explore alternatives for people like me who like to share my views and engage in open and hopefully civil communications:

1. Many people do not want to be part of a platform that is becoming increasingly toxic.
2. If the rate of change accelerates, some have suggested that the future of Twitter itself is at risk and it may be prudent to have some redundancy in your social media plans if so.

Regarding the second point, the changes outlined above have the potential to generate knock-on feedback loops, as highlighted by people on Twitter and elsewhere. So where to go?

There are lots of alternatives but Mastodon is mature, relatively well-known and similar to Twitter in many ways (La Cava, Greco, and Tagarelli 2022). It has more than 30k stars on GitHub: much loved by users and developers alike.

Furthermore, #Mastodon is trending on Twitter itself, signalling a group dynamic element. Many of the people I rate and follow on Twitter are already on Mastodon, including @underdarkGIS, @djnavarro and @UrbanDemog.

Fun theory on this: Elon Musk is genuinely committed to creating a more open world. But his approach to achieving it is not to “‘Open Source’ Twitter’s Algorithms” as stated in a Bloomberg article earlier this year, but to deliberately sabotage proprietary platforms so that open source alternatives flourish. That may seem far fetched but if it was Elon’s true aim he has been successful: Mastodon already has a strong developer community (Mastodon’s codebase has almost 700 contributors from around the world) and could represent a short cut to ensuring that the dominant platforms people use are open source.

As outlined in a post by the people who run the fosstodon instance, sign-ups to the server are booming.

So, assuming you want to join, let’s take a dive in!

Which Mastodon server to join?

Short answer: it’s up to you. Fosstodon.org seems like a good option to me. The choice of Mastodon server is covered in a blog post by Danielle Navarro, who points out advantages of the more decentralised model compared with Twitter’s centralised model:

First off, mastodon is not a single application: it’s a distributed network of servers that all talk to each other using a shared protocol. If two servers talk to each other they are said to be “federated” with one another, and the network as a whole is referred to as the “fediverse”.

There are many different servers out there that are independently running mastodon: these are called mastodon instances. You can sign up for an account at one or more of these servers. The most popular instance is mastodon.social, but for reasons I’ll talk about in a moment this might not be the best choice for you!

That post goes on to highlight a key benefit of the approach: it does not matter that much which your primary account is because you can still interact with federated instances. You can choose from a range of servers, some of which are described later in the post based on your interests. Fosstodon.org is a reasonable choice for people interested in free and open source software and data science but there are many other good options, including:

• fosstodon.org one of the big ones with 30k+ users including my Geocomputation with R colleague Jakub Nowosad and Anita Graser
• masto.pt, a server for Portuguese communities with 3k+ users, home to Rosa Felix
• mapstodon.space, a new server for mapping, currently with just over 100 users including Alex Singleton
• mastodon.online, home to many people including Antonio Paez and Cait Robinson
• sciences.social, home to Barney Harris
• toot.community, home of Rafael Pereira
• mastodon.green, home to Ben Anderson

A really cool thing about this list, other than the diversity, is that people can and do follow each other between those instances. To see a list of the top instances at instances.social.

Getting set-up

The sign-up process is fairly simple: choose a server from and click to sign up! There are detailed instructions in many articles online, one of the best I have found is published by PC Mag.

Crossposting with moa.party

Crossposting is the process of automatically posting from one platform or account to one or more other platforms or accounts. Although I like what I’ve seen of Mastodon so far, no online system can be guaranteed to see 100% uptime. Indeed, the great people running fosstodon.org outlined issues that they have faced already:

This [surge of interest in alternatives to Twitter] effectively resulted in a sustained DDoS that lasted for around 36 hours. As you can imagine, that was a lot of fun for myself and the team.

As shown in the screenshot below, I even experienced some temporary downtime shortly after joining (and getting lots of boosts):

In this context, redundancy is good. Furthermore, while I may have the capacity to set-up a new social media account, I cannot guarantee that others will. So, unlike the author of the PC Mag article on migrating, I plan to keep my Twitter account alive. If you do, crossposting can be good for various reasons, in particular for accessibility: making work widely available to the public is one way to maximise the chances of your work having positive impact beyond your immediate sphere of influence.

So how to crosspost? Simple answer: enter moa.party.

As shown in the screenshot below, you can choose the options that are appropriate for you.

After that it simply worked! I ‘tooted’ from my Masterdon account and, after a minute or so, the same content appeared on Twitter. Magic thanks to moa.party!

I cannot guarantee that it will always work and I don’t know how scalable the service is. But for now I’m stoked that not only have I set-up a new social media account — on a platform that is based on open source software and community values — but also have a way to get my work out to as many people as possible. The two versions of the message below show that Masterdon toots are very similar to Twitter tweets, spot the difference!

vs..

OK here goes, first attempt at ‘tooting’ from my new \#mastodon account 🐘 📢 Topic: georectifying \#CAD drawing of housing development and geodigitizing \#osm data. Appropriate topic for \#fosstodon? Hope so, more soon! pic.twitter.com/fwnHJeY9Mj

— @robinlovelace November 5, 2022

Even if Twitter does become the MySpace of the 2020s, it is unlikely to happen overnight. And the platform has many great people, despite some of the recent changes outlined above. Let’s see what happens. If nothing else, having alternatives like Mastodon allows you to choose to interact in the virtual spaces where you feel most at home.

Appropriate social media platforms for the 2020s

That brings me to my final thoughts on the topic, which informed the title of this article. What are the desirable attributes of social media platforms in the 2020s? A bit like with software, people choosing social media platforms today are lucky simply because they can choose. When MySpace, Facebook and other sites first launched they had ‘first mover’ advantage. There are now many options. I really liked (and still do like, on a good day) Twitter because, unlike Facebook or LinkedIn, it had (and still has) a publicly-accessible-by-default policy. That encourages sharing, transparency and collaboration across borders and cultures.

Unlike Twitter, however, many of the fastest-growing social media platforms of the 2020s are open source. This discourages (but may not completely prevent) abuses by the owners of the platforms for financial incentives. The excesses of large for-profit social media platforms are well documented in documentaries like The Social Dilemma. The fact that emerging social media platforms such as Mastodon are open source is not just important for transparency and data security: it encourages innovation and diversity. This applies to open source software in general, encouraging participation in knowledge creation (Dhir and Dhir 2017).

When the underlying technology is open, people are encouraged to set-up their own instances, spreading the compute, moderation, and content creation work widely. If you don’t like a particular space in this ‘fediverse’ world, you can simply move. And migrating from one Mastodon server to another is easy, as I have found, following in the footsteps of Anita Graser: I set-up a Mastodon instance at https://mastodon.social/@robinlovelace years ago out of interest. That now redirects to my current Mastodon home @robinlovelace@fosstodon.org.

In summary, there are good reasons for exploring alternatives to centralised for-profit social media platforms. Mastodon is relatively easy to set up and seems to enable two-way cross-posting with Twitter. And social media platforms that are appropriate for the 2020s are:

• ✅ Open source
• ✅ Easy to migrate to and from different servers without lock-in
• ✅ Fun to be a part of

Mastodon ticks all those boxes and I recommend checking it out. I also recommend donating to the free and open social media platform of your choice. I have just set-up a recurring donation to fosstodon.org via their Patreon page. If that helps the ad-free platform provide seemingly good-quality content to its users (while minimising downtime), it’s money well spent for me. I’m glad to have received the message below!

Like most people I am new to this space and am learning. So any tips/corrections that people have, like the reply to my first toot below from @rudolf@fosstodon.org (thanks for the tip Rudolf!), are welcome on Mastodon, Twitter, comments below or anywhere.

As all social media platform pioneers know, the most valuable thing you can contribute is good content, however. I plan to post high quality content online, increasing the signal-to-noise ratio in online comms, countering disinformation, and building strong communities in the process. I hope you will too. Be the change!

References

First solo authored paper

Collaborating with others is one of the most rewarding parts of academic research. I have published over two dozen papers with many people over the years, and 5 fully peer reviewed papers in 2021, which can be seen on my website and here.

One of the most rewarding papers, however, was on that I authored solo. In it, I review 25 open source tools for geographic analysis in transport planning and discuss the future potential of open source software for guiding effective policy interventions. The paper has received some attention online, as can be seen in the Tweet below, and not just from the academic research community: experienced transport modeller Tom Van Vuren, for example, picked-up on the paper and we hope to take the conversation about uptake of open and transparent software for transport modelling forward in 2022.

Stoked to announce that the new #openaccess paper "Open source tools for geographic analysis in transport planning" has just been published 🎉

Paper: https://t.co/kZy2kkyjDy
Code: https://t.co/yOFtKbBbYc
My first ever solo-authored paper! pic.twitter.com/3fHSbP0UTG

— Robin Lovelace (@robinlovelace) January 19, 2021

I also hope to do more research that is policy relevant, leading to the next topic in my reflections on the year just gone by.

New fellowship: placement in the No. 10 Data Science team (10DS)

An exciting new project started for me in 2021, as I was awarded an 18 month ‘10DS Fellowship’ to help central government make better use data science to support evidence-based decision-making.

Really excited about this: for the next ~18 months I will be working with the @10DowningStreet Data Science team, @DataSciCampus, @ESRC and others on the new 10DS Fellowship. Looking forward to working with Federico, Alex and everyone on data science in government 🚀 https://t.co/h7skJ0rbwA

— Robin Lovelace (@robinlovelace) November 18, 2021

The responses replying to that tweet from acquaintances working in related fields, plus emails from colleagues at the University of Leeds, provide a strong motivation to work hard to move things forward. New methods and datasets have great potential to be harnessed to help tackle entrenched challenges in the 21^st century. I hope my background and skills will enable me to make a positive difference.

Watch this space for projects that I end up working on as part of this fellowship (that can be published after clearance from No. 10). The placement phase of the fellowship starts in January, and it looks like there will be many exciting opportunities for high impact research. I hope to document lessons learned from the experience in articles for the organisations supporting the fellowship: the ESRC, ADR UK, the Alan Turing Institute (where I was also awarded a fellowship in 2021) and the ONS’s Data Science Campus.

Software development

In addition to research published in academic papers, I have also worked hard on software development and teaching materials. I was amazed and slightly humbled to hear that stplanr, an R package I developed to support strategical cycle network planning in the UK, had been used in research into access to potentially life-saving TB clinics in Malawi.

When a tool first built for one purpose (cycle network planning in Europe) gets used for something totally different (calculating distance to tuberculosis clinics in Africa) 🚀

Paper: https://t.co/8gaRigYXVK
Package: https://t.co/t8UC0JGjNg#RStats package {stplanr} in action! pic.twitter.com/JqBgv7CE9r

— Robin Lovelace (@robinlovelace) September 10, 2021

What better motivation could there be for developing open source software for reproducible data analysis and policy interventions?

Geocomputation with R Second Edition

Perhaps less exciting than starting new projects is ‘finishing’ existing ones. In this case the project is the popular book Geocomputation with R which, alongside other open source teaching materials of the type published on bookdown.org, helps reduce financial barriers to accessing textbooks by providing teaching content that anyone can read.

Jakub, Jannes and I have signed a contract to work on a second edition and we’ve made good progress, as shown below (one new year’s resolution: publish the 2nd edition!).

🏗️ Geocomputation with R update 🏗️ we've recently updated key parts of Chapter 7 on the key topic of coordinate reference systems (CRSs). Updated figure below, plus lots more in the chapter, shows importance of understanding them! https://t.co/pffHqeF8YT #Rstats #rspatial pic.twitter.com/DVgHU5lvn1

— Robin Lovelace (@robinlovelace) December 30, 2021

Becoming a father

The most important event in the year for me was becoming a dad.

Big personal news: @katylvlace and I have had a baby, welcome to the world Kit 🎉 Focussed on nappies and supporting the family for the next while 💚💚💚 Parenthood is amazing, makes me more committed to research for a sustainable future 🌱 Looking forward to it all! pic.twitter.com/sJJZNZZV04

— Robin Lovelace (@robinlovelace) October 14, 2021

It has been a life-changing experience: being responsible for another (little and very cute) person has certainly shifted my perspective on the world, made me prioritise my time more carefully, and made it harder to plan.

My wife Katy and I have worked closely as a team and, despite talk of the ‘baby bomb’ impacting our relationship, little Kit has brought us closer together ♥

Despite inevitable challenges, it’s been the most rewarding thing we’ve done so far, and we’re only 3 months in! Having a little one around encourages long-term thinking, and Kit has strengthened my commitment to research supporting a more sustainable world.

Just look up ☄

Having a baby around on New Year’s Eve, plus the fact that we start the new year in the midst of a pandemic, led to scaled-back plans to celebrate the arrival of 2022. To celebrate, Katy and I watched Don’t Look Up, a sci-film satirising so-far sluggish policy responses to threats identified thanks to scientific research, such as climate change.

Having heard great things about it from friends talking about it and seeing positive and provoking comments on social media, with the hashtag #DontLookUp☄, we decided to watch it as part of our NYE celebrations. The film contains the humorous yet inspiring song below by Ariana Grande.

Never thought I would hear these lyrics in a catchy pop song in 2021:

Get your head out of your ass, listen to the goddam qualified scientists

So I’m prepared for anything in the coming year. Optimistic that with hard work, our collective intelligence, empathy, and desire to collaborate on solving tricky problems will prevail. Bring on 2022!

Leeds Climate Innovation District

Leeds Climate Innovation District (LCID) is a major development in Leeds, with 500+ dwellings either in the process of being built or already being lived in (at the time of writing in April 2021). I cannot comment on the energy efficiency ratings of the buildings (although I have little doubt that they are well insulated) but in terms of its appearance and accessibility by walking and cycling it is ground-breaking. It is located on the East of the River Aire and to the West of the A61, a major road that wouldn’t seem to provide a basis for sustainable and healthy living, as shown below.

I had heard of LCID and passed by it during early stages of construction but only appreciated its scale and level and ambition today after seeing the site. In the spring sunshine it was stunning, as shown in the photo below.

My more immediate interest in the project was due to work. The LCID is one of the case study sites we used to highlight the potential for the planning system to support active travel in the ActDev project. As can be seen at actdev.cyipt.bike/lcid/, the site performs well from a sustainable transport perspective, as shown below.

The reasons why the site scores so high from a sustainable perspective perspective are simple:

• A high proportion of trips in the surrounding area are made by public transport
• Many key destinations are within walking and cycling distance

The data driven approach can hide important details on the ground that can help or hinder active travel. I was happy to see that the site had good accessibility to the key walking and cycling route to the West of the river, as shown in the photos below.

The walking and cycling provision on the LCID side of the river is good, with steps that are wide and a ramp for cycling, wheelchairs and pushchairs. The ramp onto the bridge from the East side of the river has a shallow incline, although is unnecessarily maze-like (possibly this is to reduce cycling speeds to the benefit of other users which seems sensible given the wide range of people using the path).

The waterside cycle path in East Leeds

From the LCID we proceeded southeast along waterside cycleway that follows the River Aire all the way out to Allerton Bywater, which is another new housing development that is part of the ActDev project. The route has recently been resurfaced, as described in an article by Sustrans. Having used this cycleway for more almost a decade it is hard to put into words how much better it is after the upgrade, but I will try. It is waaaaaaaaaaaay better.

Before the upgrade the path was thin, bumpy and often muddy Now it is wide, smooth and attractive. The difference can be seen in the before/after photos below (credit: Sustrans).

The surface is as smooth as tarmac on a road (smoother than most roads in Leeds, which have many potholes due to constant erosion by motor traffic) but has a rustic feel.

One issue with the route (that is already well documented and is the subject of a campaign by Leeds Cycling Campaign) is that you are forced onto steps to go from Skelton Grange Road bridge onto a cycleway located between the River Aire on the East and the Aire and Calder Navigation canal to West, as shown below.

Aside from the steps, this section is also great to ride on and is part of the Trans Pennine Trail from Leeds to Sheffield 40 miles to the South.

The Extra service station and Temple Newsam

The most surprising aspect of the ride was the new walking and cycling infrastructure provided in and around a new service station serving the M1 motorway of all things. Despite its proximity to this huge highway, the site surrounding the Leeds Skelton Lake – M1 J45 services is outstandingly good for cycling and walking, with dedicated space inside the premises and a decent shared cycleway/pavement leading up to the M1, as shown below (note the petrol station, part of the Extra service station to the left).

Crossing the huge roads to cross the M1 was possible but not ideal: the cycle infrastructure directly next to the motorway decreased substantially in quality compared with the rest of the route.

In addition to work to improve links with the Skelton Grange Road Bridge I imagine that investment in the quality of infrastructure leading to Temple Newsam, shown in the map below, would be well spent.

The Wyke Beck Way and Leeds Cycle Superhighway 2

The Wyke Beck Way is an off road route that goes from Temple Newsam all the way to Roundhay Park, the biggest park in Leeds. We were briefly on this route leaving Temple Newsam and it was our plan to take this route from the Wykebeck Arms (shown below) back to North Leeds.

However, we decided to head East along the Leeds Cycle Superhighway 2. Less famous than the Leeds-Bradford ‘Superhighway 1’ this cycleway follows the A64 dual carriageway from Seacroft and Crossgates in the East to the city centre. Despite being (I guess) the most expensive piece of cycle infrastructure per km (with dedicated cycleways in urban areas costing ~£1k/m which is still much cheaper than motorways that cost around 50 times more m), this was also the most disappointing. The superhighway had very few people walking and cycling on it compared with the busy yet not crowded river route. It was not a particularly pleasant experience, as shown in the figure below, but was much better than cycling on the road would have been.

Where next?

After following the superhighway for a mile or so we took right onto Harehills lane. This was the longest on road section and I would not recommend it to inexperienced cyclists, certainly not the ‘8 and 80 year-olds’ and less able bodied people who we should be designing for. Still, if you are confident on a bike, it is a great route: going down onto Roundhay you have a great view of this part of the city.

In terms of where to next, I plan to try routes in different parts of the city to see how they have changed. Overall I am really impressed at the changes I’ve seen in urban infrastructure described in this blog post. I hope other changes have been equally good and that the momentum in a positive direction continues.

Links

• The GPS trace associated with the route that you can download and follow with apps such as the free and open source OsmAnd Android app: https://www.openstreetmap.org/trace/3633523/
• For more routes and information on provision for cycling in Leeds see here: https://www.leeds.gov.uk/parking-roads-and-travel/connecting-leeds-and-transforming-travel/cycling/cycle-routes
• To see a simulation of LCID in your web browser for the fun of it (motivating the visit) see here: https://actdev.cyipt.bike/lcid/routenetwork,buildings/#
• To contribute to OpenStreetMap see https://www.openstreetmap.org/fixthemap

One sentence review

Outstanding and original description of artificial intelligence (AI) in the 21st Century, the problems it could pose, and a lucid proposal for ‘provably beneficial AI’.

Review

I recently completed the book Human Compatible by Professor of Computer Science at Berkeley, Stuart Russell (Russell 2019). The first book on the subject I have read, it greatly enhanced my understanding of a complex subject and made me feel compelled to write a brief review, for the benefit of myself (I’m still consolidating my understanding of the books key ideas and find that writing them down helps) and others.

I first encountered Stuart Russell and his ideas in a podcast for The Guardian’s Science weekly in October 2019, which is still available and well worth a listen. As someone who has spent much time thinking about, using and to some extent developing technologies to support a small but influential area of human activity (Transport Planning), I am acutely aware of the potential for technologies to have really world impacts. In my case, the technology is fairly narrow and, as far as I am aware, can only be used for the definitively benign activity of cycle network planning (Lovelace et al. 2017).

But imagine you were developing a much more ambitious and generally applicable technology with the explicit objective of becoming more intelligent than humans in increasing range of areas: to develop general computer intelligence. That is the starting point of Stuart’s book, the history of the academic field of AI and the astounding progress AI researchers have made in the last couple of years.

Stuart delves into the issues that AI researchers have themselves been grappling with, starting with ‘father of AI’ Alan Turing:

Even before the birth of AI in 1956, august intellectuals were harrumphing and saying that intelligent machines were impossible. Alan Turing devoted much of his seminal 1950 paper, “Computing Machinery and Intelligence,” to refuting these arguments. Ever since, the AI community has been fending off similar claims of impossibility.

The point is that the great Alan Turing and AI researchers since then have had to debunk simplistic arguments to the effect that ‘AI is impossible’, from a range of perspectives. Paradoxically, it seems that as AI researchers have actually got closer to achieving general intelligence their willingness to discuss the wider implications have reduced (or at least not increased as one would have expected). Surprisingly, some AI researchers have even began to deny the possibility of succeeding in own area. This is despite rapid growth of the range of tasks at which machines outperform humans, from playing games such as Go to commercially valuable services such as (in some contexts) human language translation and medical diagnosis of diseases.

A striking example of this denial is from Stanford University’s prominent AI100 report, which states simply that

unlike in the movies, there is no race of superhuman robots on the horizon or probably even possible.

This is astonishing because of the context: much of the rest of the report is focussed on the impacts and benefits of AI, and progress made in the field. Russell sees this statement as reflective of a wider problem: unwillingness to engage in possible negative consequences if AI systems get out of control:

To my knowledge, this is the first time that serious AI researchers have publicly espoused the view that human-level or superhuman AI is impossible—and this in the middle of a period of extremely rapid progress in AI research, when barrier after barrier is being breached. It’s as if a group of leading cancer biologists announced that they had been fooling us all along: they’ve always known that there will never be a cure for cancer.

Prof Russell uses the book as an opportunity to call for more open-minded debate among the AI community of which he is a part and in wider society. Of course the future is uncertain, meaning we should explore a range of possible futures and act to minimise the probability of highly undesirable outcomes. In various scenarios of the future sketched out concisely (and somehow entertainingly) in the book, superintelligent AI leads to the demise of humanity.

This outcome is thankfully not guaranteed but if it at least may be possible surely it is worthy of consideration? Another common objection to debate of or research into AI risks is that the threat is far in the future so we don’t need to worry. Russell uses the metaphor of asteroid collision to debunk this argument:

For example, if we were to detect a large asteroid on course to collide with Earth in 2069, would we say it’s too soon to worry? Quite the opposite! There would be a worldwide emergency project to develop the means to counter the threat. We wouldn’t wait until 2068 to start working on a solution, because we can’t say in advance how much time is needed. Indeed, NASA’s Planetary Defense project is already working on possible solutions

The good news is that there are solutions, the subject of Chapter 7 on AI: A different approach in which Russell sets out principles for beneficial AI:

[beneficial machines are those] whose actions can be expected to achieve our objectives rather than their objectives. … The resulting approach should lead eventually to machines that present no threat to us, no matter how intelligent they are.

The approach is based on three broad principles:

1. The machine’s only objective is to maximize the realization of human preferences.

2. The machine is initially uncertain about what those preferences are.

3. The ultimate source of information about human preferences is human behavior.

More good news: recent progress in AI research and processing uncertainty help make these things possible. The result of implementing these principles could be machines that regularly check-in to find out if what they think is wanted is actually what the human wants.

To me, the approach also has another advantage it leads to humility. Even the most intelligent people change what they want from time-to-time, and it makes sense to imbue machines with this humility. Instead of telling machines “do this at any expense”, it makes sense to say something more like “try to do this, but precede with caution and check in if you have any doubts”. The resulting two-way human-machine communication could lead to more evidence-based and informed decisions.

Judging by the 169 reviews of the book on Amazon UK, many of them glowing, at the time of writing, it seems that people are taking notice of Russell’s ideas. A Brit in the USA, Russell, he has an audience over the Atlantic also (although perhaps not quite as large or vocal an audience, with 164 reviews on the American amazon.com). As an informed data scientist who has seen the benefits of digital technology I am open to the possibility that there could be unintended consequences of my work.

I sincerely hope that AI researchers are equally open to critique: even if Russell’s principles get implemented to the letter, there are still risks ranging from rogue researchers to superintelligent AIs persuading us to unleash it from our control. But if debate/research/investment follows the lines suggests in Stuart Russell’s lucid and entertaining Human compatible, the probability of catastrophic AI malfunction will decrease, raising the hope of using our ever more sophisticated technologies to free us rather than enslave us.

If we succeed in creating provably beneficial AI systems, we would eliminate the risk that we might lose control over superintelligent machines. Humanity could proceed with their development and reap the almost unimaginable benefits that would flow from the ability to wield far greater intelligence in advancing our civilization. We would be released from millennia of servitude as agricultural, industrial, and clerical robots and we would be free to make the best of life’s potential.

References

References