Air quality / pollution data

[zarino]

From Mary:

Finding usable data on air pollution is proving tricky. Global Action Plan hosts this https://www.cleanairhub.org.uk/forecasts but it is live changing data, pulling info from the government defra/met hub, which works with patterns and weather predictions. What I can't see is the data defra is using to make the predictions?

Beth shared a link to addresspollution.org, which says of its source data:

Imperial College London’s Environmental Research Group (ERG), who are the UK’s leading authority on air pollution. With their sophisticated modelling techniques and network of state of the art air pollution sensors, combined with the monitor data sourced and mapped by COPI from every UK council, ICL have provided us with a national air pollution model accurate to 20m sq.

addresspollution.org is using the latest available annualised data from 2019. When newer data becomes available, the site will be updated.

It’d be great to find an up-to-date source of air quality / pollution data per constituency.

[JuliaNgoga]

Data sets here and here

[zarino]

Mike at FoE let us know that they’ve recently published more up-to-date air quality data, "using data released by DEFRA last week and data on energy efficiency and tree cover" – worth checking whether this is any better that what we’ve currently got. Although, sadly, it’ll be England and Wales only. https://policy.friendsoftheearth.uk/download/environmental-data-political-constituency

[zarino]

Mike and Toby from FoE, and I, had a short discussion about air quality data before Christmas, and they shared how they compare LSOA air quality figures to WHO guidelines to make them easier to understand. Toby from FoE has shared their data on this, for current constituencies – I’ve put it into the shared drive.

[zarino]

I note that DEFRA now lists 2022 shapefile data on the page where we previously got 2021 data: https://uk-air.defra.gov.uk/data/pcm-data, so we could update that too, at the same time as comparing with WHO guidelines. /cc @alexander-griffen

[zarino]

@juliacush and I discussed this yesterday, in reference to https://github.com/mysociety/local-intelligence-hub/pull/520. We were trying to work out the most useful way to incorporate the WHO guidelines, to give users context on what "good" or "bad" pollution numbers look like.

It’s worth noting that only three of the pollutants available to us have comparable WHO guidelines:

Pollutant	DEFRA data units	WHO 2021 guideline	Comparable?
PM2.5	annual mean, μg/m3	annual mean, 5 μg/m3	✅
PM10	annual mean, μg/m3	annual mean, 15 μg/m3	✅
NO2	annual mean, μg/m3	annual mean, 10 μg/m3	✅
SO2	annual mean, μg/m3	24-hour mean, 40 μg/m3	❌
O3 (Ozone)	days with max 8-hr concentration greater than 120 μg/m3	10 μg/m3	❌

@juliacush and I discussed that one way we could ground the constituency numbers against the WHO guidelines (in a numeric way, that can be used in filters and tables, rather than just through shading) would be to calculate the ratio of the constituency’s value and the corresponding WHO guideline value, perhaps as a percentage. eg:

Constituency	PM2.5	WHO guideline	Ratio
Cardiff Central	9.17 μg/m3	5 μg/m3	183%
North Devon	5.35 μg/m3	5 μg/m3	107%
Argyll and Bute	2.53 μg/m3	5 μg/m3	51%

To keep the raw pollutant numbers accessible to specialist campaigners who want to filter by a specific number (for example, the government’s target of 10 μg/m³ of PM_2.5) these percentages could be displayed as their own separate dataset – "Air pollution: As a % of WHO health guidelines" or similar.

The question, then, becomes whether exposing three different figures for air quality (PM_2.5, PM₁₀, and NO₂) is simple enough – or whether we dare to average them together into a single (very rough) overall pollution ratio. Scientists would probably balk at the idea, but if our aim is to give non-specialists a single indicator of "is air quality a problem in this constituency" maybe it’s worth it?

[zarino]

James from Democracy Network suggested, if we need anyone to sense-check our plans for displaying air quality data, we should contact Simon Birkett at Clean Air in London.