develop suggested process for visualizing the output

[danbjoseph]

once we have a bunch of points with Green View Index (GVI) scores, we need a way to look at it 😁

sample dataset:

• Three_Rivers_Michigan_USA_points_2_images.gpkg.zip
• Three_Rivers_Michigan_USA_points_2_images.geojson.zip

ideas?

• e.g., maybe a hex-bin density map like described here: YouTube tutorial; blog post example 1; blog post example 2
• e.g., perhaps documenting how to use tippecanoe and QGIS in a workflow like the one described here: https://digital-geography.com/create-your-own-vector-basemaps-the-easy-way/
• is there a way we could have a self-contained leaflet.js map with all the data and requirements contained in a GitHub Pages?
• something else???

something better than just this...

[ioalexei]

Example of 100m hex using the sample data:

Some thoughts on this:

• Hexes seem a good way to summarise the data - H3 would be great for this and has support in Python through h3
• They are easy to show as a layer on a webmap - e.g. MapLibre or Leaflet - that can easily be embedded in a GitHub page. Since the data would be quite small and simple, I don't think you'd need tippecanoe as an intermediate step - you could just use a single geojson as the data source
• If using hexbins as output, could let the user specify:
  • Zoom level(s) to generate/summarise (with a sensible default derived from the input data extent)
  • Summary statistic to visualise - i.e. whether each polygon should be visualised using the GVI sum, mean, median, max, min, etc. Could default to mean.
  • Bounding box to generate hexes for (with a default of the bbox of the input data?)
  • Whether to filter out null cells - i.e. hex cells within the bbox but not containing any points
  • Output format for hex polygons
  • Whether to generate a Leaflet/MapLibre map as an additional output (or perhaps that should be a separate feature)

[jchboettcher]

What's the downside of just a standard heat map? Might be simpler to implement than a hex grid.

[ioalexei]

Demo/comparison of hex vs heatmap:

Aesthetically I think there's not much in it, and both are easily supported by webmapping tools.

A benefit of using a hex grid would be the ability to show attributes as labels or in pop-ups - e.g. you could label the polygon with the GVI mean, or show a popup that has the summary stats.

Could enable both and give an option to toggle between them on the webmap?

Edit: Depending on the scope of the project, hex grid potentially makes it easier to join other data - e.g. to also show population or other sociodemographic data alongside the GVI scores.

[danbjoseph]

For me, the heatmap suggest an absence of the variable in the locations with no color. That is, areas of no vegetation. Whereas, for me, the hex grid is makes me think "no data" or "not part of the analysis" for the areas outside of the hex. It would be interesting to do a user study and ask them how they interpret the different options to figure out which most accurately helps them take away the intended information. The hexbins have the added benefit of being a simple drag and drop into a GIS software while also being compatible with a web map. Not dismissing heatmap, but let's do a hexbin option first.

@ioalexei so that we have something for this final step in the workflow - would you be willing to write a script to generate a hexbin file output? (following similar conventions as your work on #29)

[ioalexei]

Sure, I can have a go next week and share something for input.

[ioalexei]

Have submitted a pull request for a first stab at this.

There are still some things that could be improved in building the webmap, like dynamically setting the hex resolution and webmap zoom level based on the extent of the dataset. Suggestions welcome on the best way to do that!

[CAB-123]

looking at this real quick, both versions so far have clean lines (street, hexmap)

the hexmap also covers the building roofs and the area between the streets, which makes sense on an aggregated level but on an "extreme technical level" there's no actual vegetation