Improve spread representation with geo-heatmap

[mullerch]

Actually, the map shows the points where reports have been done based on the number of reports (for each classification. e.g. healthy).

This has multiple downsides:

• Locations that have a lots of absolute reports of a classification are over-represented relative to those who have a low number of reports but with a high ratio. E.g. a location with 10000 healthy and 1000 sick look more sick than a location with 100 healthy and 200 sick.
• Location with very few reports are over-represented. This is due to the minimal size of the marker.
• It is hard to get an overview of the situation (marker are superposed and not very readable)

To have a better representation, I would recommend developing a heatmap of the disease. Each of the three main classifications (no disease, disease, recovered) colors the location region into 'their' color, the one with the most cases (with their specific ratio) gets the 'most' coloration. Regions with no (or very few) reporting are not represented.

It could typically look like this (or a simplified version) :

Source: https://github.com/luka1199/geo-heatmap

Main challenges here are calculation and performance. Best would be to build a leaflet layer so we can keep the filters. We are also open to other propositions though.

[mullerch]

A single point to open the popup for each location would stay on top to check the effective values.

[pksec]

I like the idea.