This article tackles a longstanding problem in transport geography and planning: how to estimate realistic driving travel times without access to proprietary data, expensive APIs, or massive GPS trace datasets. Much of the planning literature still relies on “naïve” methods (e.g., minimizing Euclidean distance, network distance, or speed-limit-based traversal time) that systematically under-predict real-world travel times. At the other extreme, state-of-the-art models in computer science and transportation engineering can achieve really good accuracy, but often require billions of observations, deep learning models, and massive computational resources and capacity. We argue that planners and applied researchers need a middle ground:

• a method that uses free, open data
• runs on ordinary hardware
• and substantially improves accuracy over old naïve approaches

Using Los Angeles as a case study, we combine OpenStreetMap data on street networks, speed limits, traffic controls, and turns with a small training sample of empirical travel times from the Google Routes API. We first compute naïve shortest-path travel times, then train an interpretable random forest model to predict travel time as a function of naïve time plus turn counts and traffic controls encountered along each route. Whereas the naïve model under-predicts travel time by over three minutes on average, our model’s predictions differ from Google’s by just 0.34 seconds on average and achieve an out-of-sample MAPE of 8.4%—comparable to far more data-intensive approaches. Robustness checks with new network data yield similar performance, and SHAP analysis supports the model’s theoretical soundness.

Our goal is not to replace state-of-the-art congested travel models, but to equip less-resourced planners, scholars, and community advocates with a free, open, and accurate tool for accessibility analysis, scenario planning, and evidence-based interventions when resources are limited.

From the abstract:

Travel time prediction is central to transport geography and planning’s accessibility analyses, sustainable transportation infrastructure provision, and active transportation interventions. However, calculating accurate travel times, especially for driving, requires either extensive technical capacity and bespoke data, or resources like the Google Maps API that quickly become prohibitively expensive to analyze thousands or millions of trips necessary for metropolitan-scale analyses. Such obstacles particularly challenge less-resourced researchers, practitioners, and community advocates. This article argues that a middle-ground is needed to provide reasonably accurate travel time predictions without extensive data or computing requirements. It introduces a free, open-source minimally-congested driving time prediction model with minimal cost, data, and computational requirements. It trains and tests this model using the Los Angeles, California urban area as a case study by calculating naïve travel times from open data then developing a random forest model to predict travel times as a function of those naïve times plus open data on turns and traffic controls. Validation shows that this interpretable machine learning method offers a superior middle-ground technique that balances reasonable accuracy with minimal resource requirements.

On a side note, my co-author and doctoral advisee Wendy Zhou was also awarded the 2026 Tiebout Prize in Regional Science at the WRSA conference in Santa Fe this week. Congrats Wendy!

For more, check out our article at IJGIS or via the open-access pre-print.

Urban sustainability is key to planetary health. Yet few cities have measurable policy standards and targets to actually build healthier and more sustainable cities, and their health-supportive built environment features are often inadequate or inequitably distributed. Can residents sustainably access their daily living needs? A critical step toward making cities healthier and more sustainable is to empower policymakers and urban planners to consistently measure neighborhood-level spatial indicators of the built environment.

At GOHSC, we measure spatial and policy indicators of healthy and sustainable urban planning practice in partnership with a network of local collaborators in cities around the world. Our analytics let non-technical users load and transform their own local data to automatically calculate a set of indicators and generate a report and scorecard infographic for advocacy work. Over the past year, local planners, mayors’ offices, citizen advocates, and international organizations have used this software to measure, benchmark, and monitor cities’ progress toward global health and sustainability goals. This supports evidence-informed interventions anywhere from the neighborhood scale to the metropolitan scale. Our software’s accessibility and rigor unlock the ability to effect urgently needed change while reducing longstanding barriers to participation, particularly in under-resourced cities and countries where better planning for health and sustainability are most needed.

I’m really proud of this group. I’ve had the pleasure of serving on our executive committee and as the spatial team co-lead since its inception. We’re a member organization of the UN’s Global Urban Observatory Network and we published a series of articles in The Lancet Global Health a couple years ago about our work to date.

Sincere thanks to the hundreds of researchers and practitioners in hundreds of cities around the world who have joined our collaboration network over the past few years. And thank you to my colleagues on the executive committee: it is truly a pleasure working with you.

I’m organizing a session with Elizabeth Delmelle that asks: how can urban data science and spatial analytics meaningfully address cities’ “wicked problems” instead of just reaching for this year’s shiny new tool or the easy data set we can just grab and analyze?

Cities today face challenges that feel intractable. Despite significant progress in urban data science and spatial analytics over the past 20 years, the problems tackled by this scholarship too often prioritize the “low-hanging fruit” publication pipeline of letting easily available data and trendy methods determine the research question rather than confronting urgent but arduous challenges. Methods-driven research that gestures toward real-world applications frequently features policy implications wholly detached from the political constraints and implementation realities that practitioners must navigate.

We seek papers (theoretical, empirical, or viewpoint/commentary) that challenge and critique this status quo. This is an opportunity to experiment, try new things, attempt something hard that may not work, present null findings and research failures – as long as you propose something ambitious to use spatial methods to meaningfully improve urban living.

Submissions should use real-world problems to drive the methods (rather than letting preferred methods determine the research problem) in urban data science and spatial analytics. Example topics could include:

• spatial data storytelling for societal change
• data visualization to foster urban political conversations in an era of polarization
• evidence-based communication to persuade local stakeholders for the public good
• changing people’s minds about a crisis with frozen battlelines, such as climate change
• tangibly attenuating an urban wicked problem like homelessness or unaffordability
• shifting negative public narratives around cities that run counter to empirical evidence
• communicating the value of planning, funding, and infrastructure in an era of budget crises and defunding
• lessening (rather than merely describing) inequity in deeply segregated and unequal cities
• anything else that ambitiously tries to address a big problem facing cities through the tools of urban data science and spatial analytics that has the potential to produce meaningful and plausible change

Where/When : this session will be in-person at the American Association of Geographers Annual Meeting, March 17-21, 2026, San Francisco, California.

How to submit : submit your abstract to AAG by its Oct 30 deadline. Then email your abstract and AAG PIN to both Geoff (boeing at usc dot edu) and Elizabeth (delmelle at design dot upenn dot edu) by Nov 1 for us to consider for the session. Session participants will be notified by Nov 20.

We argue that a distinguishing feature of urban street networks, which makes them unique compared to other spatial networks, is their extreme betweenness centrality heterogeneity. In plain English that means that street networks are particularly prone to chokepoints: network nodes on which a disproportionately high number of shortest paths depend. Theoretically, we can explain this as a street network that started as a backbone road, then grew and filled in as the area around it urbanized.

Building on this idea, we propose a generative model of urban street networks - that is, a model that generates street networks that reproduce this distinguishing feature. It turns out that most models don’t! Our proposed model starts with a minimum spanning tree (the initial backbone) then adds edges iteratively (the subsequent urbanization) to match empirical degree distributions. Our model, implemented in Python, reproduces key empirical characteristics well.

From the abstract:

Analyzing 9000 urban areas’ street networks, we identify properties, including extreme betweenness centrality heterogeneity, that typical spatial network models fail to explain. Accordingly we propose a universal, parsimonious, generative model based on a two-step mechanism that begins with a spanning tree as a backbone then iteratively adds edges to match empirical degree distributions. Controlled by a single parameter representing lattice-equivalent node density, it accurately reproduces key universal properties to bridge the gap between empirical observations and generative models.

For more, check out the article at PRL or the open-access arXiv pre-print.

Most real-world objects belong to fuzzy categories, resulting in subjective decisions about what to include or exclude from counts. Yet this complexity is often obscured by a superficial impression that counting is easy to do because its mechanics seem easy to understand. After all, everyone learns to count in kindergarten by simply enumerating the elements in a set. But counting is hard because defining that set and identifying its members are often nontrivial tasks. Many of the world’s most important analytics rely far less on flashy data science techniques than they do on counting things well and justifying those counts effectively.

Street intersection counts and densities are ubiquitous measures in transportation geography and planning. However, typical street network data and typical street network analysis tools can substantially overcount them. This article explains the 3 main reasons why this happens and presents solutions to each.

Street intersections, particularly the complex kind common in modern car- centric urban areas, are fuzzy objects for which most data sources do not provide a simple 1:1 representation. This results in spatial uncertainty due to data challenges in representing network nonplanarity, intersection complexity, and curve digitization. Essentially all data sources suffer from at least 1 of these problems due to difficulties representing divided roads, slip lanes, roundabouts, interchanges, complex turning lanes, etc. If unaddressed, my assessment shows that typical intersection counts (and downstream densities) would be overestimated by >14%, but very unevenly so in different parts of the world. This bias’s extreme heterogeneity particularly hinders comparative urban analytics.

Mitigating these 3 problems is a project I’ve been iteratively refining for the past decade. It was a central focus of my dissertation and a key motivation for originally developing OSMnx. This article presents OSMnx’s algorithms to automatically simplify spatial graphs of urban street networks—via edge simplification and node consolidation—resulting in faster parsimonious models and more accurate network measures like intersection counts and densities, street segment lengths, and node degrees. These algorithms’ information compression drastically improves downstream graph analytics’ memory and runtime efficiency, boosting analytical tractability without loss of model fidelity.

Counting is hard, but we can make it a little easier by using better models. For more, check out the open-access article.

But wait, there’s more! I also discuss many lessons learned over the past decade in geospatial software development, including:

• what makes a good API, and why is it so hard for academics to make one
• how your development pipeline and continuous integration can make or break your quality of life as an open-source developer
• dependency ecosystems and the fine line between dependency heaven and dependency hell
• why reusable geospatial software is so important for open science, and how we can advance it

All of these lessons have become central to the work my RAs do in the Urban Data Lab at USC. They’re not always easy, but they make a clear improvement in code quality, clarity, and reusability that directly impacts our downstream empirical analyses and scientific theorizing.

From the abstract:

OSMnx is a Python package for downloading, modeling, analyzing, and visualizing urban networks and any other geospatial features from OpenStreetMap data. A large and growing body of literature uses it to conduct scientific studies across the disciplines of geography, urban planning, transport engineering, computer science, and others. The OSMnx project has recently developed and implemented many new features, modeling capabilities, and analytical methods. The package now encompasses substantially more functionality than was previously documented in the literature. This article introduces OSMnx’s modern capabilities, usage, and design—in addition to the scientific theory and logic underlying them. It shares lessons learned in geospatial software development and reflects on open science’s implications for urban modeling and analysis.

This year will mark the 10th anniversary of my work on the OSMnx project. It recently reached version 2.0 with a slew of new features and enhancements. If you haven’t used it before, OSMnx is a Python package to easily download, model, analyze, and visualize street networks and any other geospatial features from OpenStreetMap. You can download and model walking, driving, or biking networks with a single line of code then quickly analyze and visualize them. You can just as easily work with urban amenities/points of interest, building footprints, transit stops, elevation data, street orientations, speed/travel time, and routing.

For more, check out the article at Geographical Analysis.

Thank you to everyone who supported me along the way.

This year will mark the 10th anniversary of my work on the OSMnx project. It recently reached version 2.0 with a slew of new features and enhancements. If you haven’t used it before, OSMnx is a Python package to easily download, model, analyze, and visualize street networks and any other geospatial features from OpenStreetMap. You can download and model walking, driving, or biking networks with a single line of code then quickly analyze and visualize them. You can just as easily work with urban amenities/points of interest, building footprints, transit stops, elevation data, street orientations, speed/travel time, and routing.

If you’re interested in this tool, you can read more about it here.

From the abstract:

Measuring and monitoring progress towards achieving healthy, equitable and sustainable cities is a priority for planners, policymakers and researchers in diverse contexts globally. Yet data collection, analysis, visualisation and reporting on policy and spatial indicators involve specialised knowledge, skills, and collaboration across disciplines. Integrated open-source tools for calculating and communicating urban indicators for diverse urban contexts are needed, which provide the multiple streams of evidence required to influence policy agendas and enable local changes towards healthier and more sustainable cities. This paper reports on the development of open-source software for planning, analysis and generation of data, maps and reports on policy and spatial indicators of urban design and transport features for healthy and sustainable cities. We engaged a collaborative network of researchers and practitioners from diverse geographic contexts through an online survey and workshops, to understand and progressively meet their requirements for policy and spatial indicators. We outline our framework for action research-informed open-source software development and discuss benefits and challenges of this approach. The resulting Global Healthy and Sustainable City Indicators software is designed to meet the needs of researchers, planners, policy makers and community advocates in diverse settings for planning, calculating and disseminating policy and spatial urban indicators.

For more, check out the article.

From the abstract:

Surfacic networks are structures built upon a 2D manifold. Many systems, including transportation networks and various urban networks, fall into this category. The fluctuations of node elevations imply significant deviations from typical plane networks and require specific tools to understand their impact. Here, we present such tools, including lazy paths that minimize elevation differences, graph arduousness which measures the tiring nature of shortest paths (SPs), and the excess effort, which characterizes positive elevation variations along SPs. We illustrate these measures using toy models of surfacic networks and empirically examine pedestrian networks in selected cities. Specifically, we examine how changes in elevation affect the spatial distribution of betweenness centrality. We also demonstrate that the excess effort follows a nontrivial power law distribution, with an exponent that is not universal, which illustrates that there is a significant probability of encountering steep slopes along SPs, regardless of the elevation difference between the starting point and the destination. These findings highlight the significance of elevation fluctuations in shaping network characteristics. Surfacic networks offer a promising framework for comprehensively analyzing and modeling complex systems that are situated on or constrained to a surface environment.

For more, check out the article.