Robin Lovelace

I work at the University of Leeds as Associate Professor of Transport Data Science. I am the Lead Developer of a national active travel mapping project to design strategic cycle networks nationwide (see www.pct.bike). I have written open source software packages and books including Geocomputation with R.

The speaker's profile picture

Sessions

10-01
13:00
30min
Zonebuilders: cross-platform and language-agnostic tools for generating zoning systems for urban analysis and modelling
Dustin Carlino, Robin Lovelace

Zones are key building blocks used for analysis and creating models (mental and statistical) of urban and environmental systems.
Used in a range of fields from biodiversity assessment to transport planning, spatially contiguous areal units break-up continuous space into discrete chunks.
Many methods rely on good zoning systems, including origin-destination analysis, geographically weighted regression, and choropleth visualisation.

Open access administrative boundaries are increasingly available through national databases and OpenStreetMap but are often inappropriate to geographic research, analysis and map making needs, being often: based on arbitrary factors; inconsistent between different cities/regions; and of highly variable sizes and shapes.

This talk outlines an approach to tackle these problems: tools that can auto-generate zones based on minimal input data.
We propose cross-platform and language agnostic implementations to enable a diverse range of people to generate bespoke zoning systems for their needs based on the understanding that accessibility, flexibility and extensibility are key to usability.
We also demonstrate working tools that take a step in this direction which at the time of writing include:

  • a core library written in Rust with small and fast binaries available for all major operating systems
  • an R package (published on the Comprehensive R Archive Network, CRAN) that also enables visualisation of zoning systems

We plan to create a Python Package, a QGIS plugin and web user interface based on the core library and welcome suggestions and contributions via our GitHub organization: https://github.com/zonebuilders.
Based on the experience of developing these tools, we will discuss next steps towards accessible and flexible zone building tools and language/platform agnostic tools for geospatial work in general.

We conclude that the approach, based on low-level and easy-to-distribute tools that can be used in multiple free and open source frameworks, could be applied to other domains and help join diverse communities (e.g. based on R, Python or QGIS) through use of shared low-level, cross-platform and future-proof implementations.

The source code underlying the approach can be found at https://github.com/zonebuilders

Software
Puerto IguazĂș