geo-street-talk-global

Project lead: David Marulli (@dmarulli | david@argolabs.org)

Primary developer: @YukunVVan

Orgs: Streets Data Collaborative | ARGO

High-level Description

Imagine looking at any non-cartographical data visualization or UI component (tables, graphs, text, etc.) that aims to display, say, characteristics of the worst potholes in the city. To represent the location, we could display a coordinate (e.g. 40.7217267,-73.9870392), but without looking at a map, where is this?

One could imagine instead displaying an address (i.e. 263 E Houston St), but, again, without looking at a map, even to a native New Yorker like myself, I don't know immediately know if this is on the east or west side of the city.

Building on a solid foundation, the deliverable of this project will programmatically output more immediately meaningful descriptions (i.e. "Houston Street between Avenue B and Avenue C").

As folks who are in the data weeds day-to-day, it's important to not lose sight of higher-level user experience.

Approach

OSMnx is a python package that allows one to pull street networks for cities around the world. geopandas is a python package for processing geospatial data. Using these two packages as our foundation, we should be able to create a tool that recieves a lat/lng pair and returns a conversational string, such as "Houston Street between Avenue B and Avenue C". Below are the major steps we follow:

Use OSMnx to get a street graph
Given a lat/lng point, query for nearest street segment
Get to/from node IDs of street segment, and query for intersecting streets
Compare street names of intersecting streets with original segment to determine to/from streets