2024|16: health boundaries, ESRI dashboards and climate risk maps
Wrangling health board boundaries
This week we had a request from a colleague for a set of health board boundaries he could use in a PowerBI Shape Map. This ended up being slightly trickier than I expected — health board boundaries are published separately for each of the home nations, so step one was to round the all up from the official sources.
Then, I merged them into a single file in QGIS and saved them to a GeoJSON. The first output came in at a 50MB which seemed like it would slow down PowerBI when loaded — so I ran QGIS’ new area-based simplification algorithm to get a 1.2MB file.
However, the Shape Map tool takes TopoJSON format as an input, and QGIS doesn’t export to TopoJSON. Instead, I ran the file through geojson.io, which can export as TopoJSON. Despite taking more steps than I expected, the end result worked with Shape Map in PowerBI, and our colleague can now use the map to filter his data on their dashboard.
ArcGIS Dashboards
Based on another request, I spend some time exploring ArcGIS Online’s dashboarding capabilities. It’s not as fully featured or as intuitive as PowerBI, but unlike PowerBI we have the ability to publish a AGOL dashboard outside the organisation. This means, for example, that we can make a dashboard and embed it on a public website, which at the moment we can’t do with PowerBI dashboards because of the way permissions are set up in Sharepoint/Office 365.
It took quite a lot of experimentation to get a dashboard that was useable and getting filters working takes a lot of clicking around various settings, but the demo dashboard was passable. In the end, the requestor decided they didn’t think a dashboard was right for what they want to show. Still, at least I’ll have a better idea of what’s possible with AGOL dashboards the next time they come up.
Climate risk maps
Another request that came in was for a series of maps based on data from the UK Climate Risk Indicators website. I’ll continue working on this next week but for now I spent some time preparing the data — downloading the relevant data from the website and converting from GeoJSON to GeoPackage so that when I do come to use it it will be quicker to work with.