Urban-Analytics-Technology-Platform/demoland-project

Indicator of Pollution/emissions

Closed this issue · 5 comments

This issue is meant to help tracking the process of searching, selecting and generating the relevant explanatory variables (land use features) for modelling emissions as a key indicator.

After discussing this w/ Ali Ford (Newcastle), I'm starting to veer towards using some measure of air quality/pollution, relying on the Urban Observatory. This would take it away from the emissions focus, but would have other benefits.

Air quality data for UK

UK map

Data gov uk

general information here https://www.gov.uk/government/collections/air-quality-and-emissions-statistics

I was exploring the data from the urban observatory and there's a bit of a coverage issue. What looks like a decent mesh of sensors here https://newcastle.urbanobservatory.ac.uk turns out to be less impressive when you filter for those measuring air quality. You can check the map on the bottom of this notebook to see for yourself. Notice the boundary of the area of interest in black. We would get in a lot of trouble trying to interpolate anything south of the river.

One idea you may have discussed (or partially).

UK AIR data (linked above) contains 1km grid data for selected emissions. If we want to combine them to a single air quality index, we may build the expert model based on the methodology EEA is using for the European air quality index described here. That is based on the relative risks associated to short-term exposure to PM2.5, O3 and NO2, as defined by the World Health Organization in its report on the Health Risks of Air Pollution in Europe project (HRAPIE project report). That would have a direct interpretation regarding health hazards, better than some raw metrics.

I tried to do the derived index as mentioned above, it is shown and documented here (rendering from my branch, linking nbviewer to see leaflet map)