azavea/world-bank-rai-toolkit

Enhancements for RAI toolkit for July 2021 contract

world-bank-rai-toolkit

Enhancements for RAI toolkit for July 2021 contract

• Country Road Inventory Data

This project seeks to bridge information from national road inventory systems with road network geometries from OpenStreetMap. This would allow using the authoritative attributes present in road inventory systems to assign road quality attributes to geo-referenced road network features. This enables accurate calculation of UN SDG 9.1.1

This project is focused on extracting road segments from OpenStreetMap given high level description like:

• RN-01 between SALCAJA and QUETZALTENANGO with length of 7.05 KM in Guatemala
• RD-ZAC-03-01 from ZACAPA until intersection with CA-10 with length of 1.9 KM in Guatemala

Project assume the following:

• Process is only possible if relevant geometries are present in OSM
• Not all road segments may be extracted
• Exact start/end position may need manual adjustment at later time

world-bank-rai-toolkit's Issues

Inception Report

The inception report is going to say how we're going to approach the problem given the data that we have been given.

• What datasets are we discounting because they're not sufficient
• What datasets we're using
• What is the general approach
• How it may tie in with future development of RAI toolkit

Evaluate Paraguay Data

WB sent us HDM4 export of Paraguay data.
This task is to evaluate the export and see how it relates to project goals.

Because HDM4 is actually a tool used by WorldBank using it as an input could provide nice standard place to start.

• Can we read it?
• Can we extract any information that can be joined with OSM?
• Are there any blockers to using this format?

Query Overpass for Start/End Points

Some (most?) stretch descriptions are cities.

Our first step should be to use Overpass API to try to find them:

• instances: https://wiki.openstreetmap.org/wiki/Overpass_API
• Overpass has its own query language that can be used to craft a complex query
  • query by exact name may be helpful here
  • Note that cities may be points (nodes) or polygons (ways/relations)
• Overpass Turbo can be used to iterate on query type

There are two libraries that can be used to query overpass:

• https://github.com/mocnik-science/osm-python-tools
• https://pypi.org/project/overpy/

This is good point to poke at them and pick the one that seems to work best

Desired output:

• a bit of python code that queries for start/end cities
• list of cities that couldn't be located from the Guatemala invetory

Categorize Route Descriptions

We're going to use the layout of Guatemala road inventory as our model for stretch description. We're going to assume that same or similar standard can be imposed on other inventory lists.

This task is to categorize the types of stretch description present in: https://docs.google.com/spreadsheets/d/1AT6-ncQ-Fz_iDceg0mFqt2JRjChnr5O_RZi3IR19xlo/edit?usp=drive_web&ouid=100776733618911423669
.. by attempting to parse them

Its clear there are several

• SALCAJA - QUETZALTENANGO is clearly point to point
• CA-10 - ZACAPA appears point to intersection

What are others? Manual inspection of instances of each new type will be required to infer their meaning.
We're going to select a subset of representation for this project and note those left as future work.

Desired result:

• some parsing script
• list of types parsed succesfully
• list/sample of stretch descriptions ignored.

Query Overpass for Road Segment

It should be possible to query Overpass for road segment near our start/end point with name given. Related Google

This will require some experimentation but it should be possible. Fallback is to extract full road network for the country and work on it as a single FeatureCollection. That will be more time intensive, so first lest try to use overpass for this.

Desired output:

• Several instances of query that show it's possible to query road segment by name and approximate location from Guatemala dataset.