Change data loading behavior

[emilyhobson]

• Change data loading behavior. Right now we're loading and showing all data from a remote Carto layer, we would need to figure out a different approach. Most likely it would be a combination of clustering and viewport bounds limiting.

[almccon]

I worry that clustering will be a non-starter for the clients. Viewport bounds limiting is probably fine. But with either of those approaches, I wonder if the development effort would be similar if we just decide to use a GL map (Tangram or MapboxGL) instead. Either of those should handle 20K points just fine.

My main concern with that approach would be to check what browsers we need to support, and how well they handle GL.

[almccon]

Actually, there are two issues here: one is map performance, and one is data loading performance.

Doing viewport bounds limiting would improve the data loading performance, because we'd be loading a smaller number of buildings (except when the viewport contains the entire city, which is the default). This would also help the performance of the comparison list (#19) without even changing the code there, just because the number of buildings would be smaller (sometimes).

Doing clustering, however, wouldn't reduce the data loading time and wouldn't reduce the number of buildings. So that's more of a map performance issue. Creating a separate issue to track that: #22.

[sconnelley]

Clustering wouldn't reduce data loading times but it would visually make the map look better. Viewpoint limiting would be negated when the map is zoomed out to the city extent. So if we're talking a city has 20k points, that would be visual diarrhea

I think we need to run some tests on a dummy dataset containing 20k points, to see truly the choke points.

@almccon -- could you create the dummy dataset?

Also, since we have issue #22, can we close this one?