Choropleth experiment using Altair and ipywidgets

[edwardchalstrey1]

Take the Choropleth charts in Altair (from James Trip) and add widgets for toggling between different colour-binning possibilities, plus appropriate metavisualisations based on the ones by Greg McInerny.

• Tidy code (include chart function?)
• Make the extent a double slider?
• Can we have an extra subplot(s) with ipywidgets that show the binned data e.g. histogram - "metavisualisation", this is what the "WAYS" code does
• Choropleth example using Plotly #16
• Can we get all the options for scale and other params?
  • Ed note: this may well be impossible to do in an automated way, so it's possible that part of our "WAYS Workflow" could just say: "Look up the accepted values for your function online and add them to a list that gets used by Jupyter widgets dropdown...." or something
• Select different dates slider?
• Decorator experiment #15

[edwardchalstrey1]

A second example could be choosing the colours or positions (or positioning algorithms?) of a network graph package

[edwardchalstrey1]

Approach taken here is to use jupyter widgets - see ways-py/notebooks/choropleth-play.ipynb

[rolyp]

@edwardchalstrey1 Renamed this task to reflect our latest understanding of the scope.

[rolyp]

@edwardchalstrey1 Imported to do list from choropleth-play.ipynb.

[rolyp]

@edwardchalstrey1 We can discuss this when we meet at 11, but I'll mention it here as well. If I remember correctly, in the example Greg showed us:

• the histogram "metavisualisation" was to the left of the choropleth, and rotated anticlockwise by 90 degrees (so the x-axis/bins run vertically and the y-axis/record count runs horizontally to the left)
• the histograms bars themselves weren't coloured; rather, the colours were used as the "bin labels" on the x-axis

So we might want to look at implementing one or both of these before the meeting on Thurs.