Matt DelSordo

At the time of writing, layers are filtered by loading in Mapnik XML as a string, compiling it to JSON, modifying the JSON, and then compiling it back to XML....

Mapnik's `vectorTile.getData()` function has a bunch of compression settings that aren't exposed. At the time of writing, I think they're set to being highest quality/largest file size, and we might...

Currently-served vector tiles are very big, it might be possible to adjust settings to restrict available zoom levels and whatnot.

Since AWS Lambda start time and overall request speed correlates directly with upload bundle size, it would be super worth it to investigate means to shrink the uploaded bundle. This...

At the time of writing, you can 100% render OSM data in Tilegarden, but you have to run it through [`osm2pgsql`](https://wiki.openstreetmap.org/wiki/Osm2pgsql) first so that Tilegarden can read it from Postgis....

Mapnik can use shapefiles/rasters as data sources with no issue, but these need to be stored somewhere if you want this to work in production (since the Lambda bundle size...

Mapbox vector tiles are currently served as a gzipped buffer, but it might be useful to be able to request them as geojson instead. Doing that is already built in...

## Overview This PR replaces the static HTML demos with those build from Pug templates, with the intent that the demos will be easier to maintain and add to in...