Recently I've been working with climate and environmental data and thinking of ways to present this data to non-scientists.
As part of the Visigoths team we explored how to present Covid-19 risks in a dashboard in our winning entry to the 4th COVID-19 hackathon organised by the UK Natural Environment Research Council.
Later we generalised these ideas to present a range of environmental risks. This post is about the next iteration of our risk finder dashboard which explores the risks from air pollution.
You can access the live dashboard here.
Air pollution is a serious problem. According to the World Health Organisation air pollution accounts for:
The WHO continues:
- 29% of all deaths and disease from lung cancer
- 17% of all deaths and disease from acute lower respiratory infection
- 24% of all deaths from stroke
- 25% of all deaths and disease from ischaemic heart disease
- 43% of all deaths and disease from chronic obstructive pulmonary disease
Pollutants with the strongest evidence for public health concern, include particulate matter (PM), ozone (O3), nitrogen dioxide (NO2) and sulphur dioxide (SO2).
To help explore the risks that air pollution presents in Europe, we can look at forecasts of the background levels of these pollutants that have been developed as part of the Copernicus Programme and implemented by the ECMWF.
These forecasts combine models for pollutant emission sources (for example, transport, industry, domestic heating) with numerical weather prediction (NWP) models that predict how pollutants will disperse over time.
The forecasts produce predictions over a grid (with each cell being approximately 11km x 11km) at hourly intervals up to 4 days ahead. Our dashboard, based on excellent open source technologies such as leaflet.js and chart.js is designed to present the forecasts at a local level but also allow the user to explore other areas using a forecast map.
We map individual pollutant levels to risk bands from (very low) 1 to 10 (very high) using the classification system devised by UK's Department for Environment, Food and Rural Affairs (DEFRA) and also adopt the same colours that DEFRA assign to each risk band. The combined air quality index is calculated as the maximum of the air quaity indices for each pollutant. The dashboard will display the combined air quality index by default, but the individual pollutants can also be selected.
The dashboard provides an advice tab to list any precautions that may need to be taken, given the forecast for the backgound air quality index at a given location and time.
The dashboard provides a quick indication of how the risk is changing over time based at the selected location in a similar way to the warming stripes graphic.
A chart showing how the combined air quality index or indivudual pollutant levels change over time at a selected location can also be viewed.
The dashboard incorporates a backend system that downloads the latest forecasts from the Atmosphere Data Store (ADS) every morning at 9.30am UTC. The forecasts are then converted to sets of sharded geojson files at three different zoom levels. These are loaded on-demand into the dashboard itself using statically served files. This makes deployment of the dashboard possible using a simple static web server.
To try it for yourself and look at the latest forecasts head over to the risk finder air pollution dashboard.
The experience on mobile could be improved and I would like to make this more accessible to the general public by borrowing some design ideas from weather forecasts, so there is still much to do. If you have any suggestions or comments you can always tweet me on @mccniall.