Modelling World and the Scottish Transport Awards (23 June 2025)

Manchester

It's been a busy few weeks! First up was Modelling World, where I had the chance to demo the Network Planning Workspace. There were some incredible talks about generating synthetic populations, building route choice models for London cyclists, and using MATSim to study per-person exposure to pollution. There was a theme of complex models needing to consider data visualization and storytelling from the start.

Demonstrating NPW in a very loud room. Photo credit Robin Lovelace.

I also had a chance to properly wander around Manchester for the first time, and really enjoy the public spaces in peak June weather.

Glasgow

Then it was off to Glasgow to attend the Scottish Transport Awards, where NPW was highly commended in the transport planning category.

Sadly, the full NPW team couldn't make it.

I had a few day to explore Glasgow by bike, which really was the best way to get around. The segregated infrastructure I sampled all felt awesome, and the canal tow-paths were gorgeous. With a Sustrans colleague, I also cycled the National Cycle Network 7 to Loch Lomond.

The NCN 7 never felt dull, constantly changing scenery.

For all of the time I spend behind a computer thinking about infrastructure quality, it's always refreshing to experience some egregious pavement parking firsthand.

Project updates

In between the travel, I've had enough time to continue improving the NPW and LTN tool. NPW has mostly been performance work, like avoiding overhead between Rust and JS that happens with Comlink's structured cloning for web workers. The LTN tool has a few new features:

You can now easily compare a neighbourhood before and after your edits (details)
You can reclassify local/main roads by snapping a route, instead of tediously tracing small lines (details)
More shortcuts are now detected correctly (details)

Finally, I've been experimenting with data in OpenStreetMap about crossings and sidewalks mapped as parallel lines. There'll be much more to say about that later, but I'm happy to report that an initial experiment to check crossing frequency works fine in areas without detailed separate sidewalks.