mapbox / mapboxgl-jupyter

Use Mapbox GL JS to visualize data in a Python Jupyter notebook

MIT License

666 stars 137 forks source link

Mapboxgl viz types roadmap #4

Open ryanbaumann opened 7 years ago

ryanbaumann commented 7 years ago

This ticket tracks the project roadmap. If you have a specific feature request, please open a new ticket and track there!

Visualization Types

[x] Circle map
[x] Graduated Circle map
[x] Clustered Circle map
[x] Heatmap
[x] Raster Tile map
[x] Image map
[x] Choropleth (fill map)
[x] Linestring map
[ ] Arc origin / destination map
[x] 3D extrusion map
[ ] Symbol map
[ ] 3D terrain / environment map
[ ] Add multiple visualization layers to one map

Utilities

[x] Export rendered viz to image
[ ] Create viz data directly from Pandas or Geopandas dataframe

Viz interaction

[ ] Hover/Highlight effects for mouseover of features
[ ] Data selection and filtering on map with a lasso, rectangle, and circle selection tool

ulemanstreaming commented 6 years ago

A strong vote for Clustered Circle maps. Dismayed to find it missing in the roadmap for 1.0.

Clustering is one of the best map visualization tricks to come along in recent years, making it possible to use large, global point data sets without prior subsampling or scale-based subsetting. Ideal for the data exploration stage of data science workflows.

rselover commented 6 years ago

@ryanbaumann - based on your profile, it looks like you've got a lot going on!

My vote is for Choropleth - do you have a sense of when Jupyter dev will get more attention?

ryanbaumann commented 6 years ago

@rselover heya!

Clusters and circles are next on the priority list. We're working on some core Mapbox GL JS for data visualization abstractions that should make the python mapboxgl library here much more supportable and performant!

cc @perrygeo

ryanbaumann commented 6 years ago

Added preliminary clustered circle viz support in https://github.com/mapbox/mapboxgl-jupyter/pull/27 cc @ulemanstreaming

rmcarthur commented 5 years ago

How relevant is this roadmap? Is this still the most important goal posts to look for?

akacarlyann commented 5 years ago

@rmcarthur I'm not sure about the general vision but this is what I reference as a volunteer contributor :). I think multi-layer maps are the highest priority. I'm actually hoping to open a PR for this soon. Hexbin visualizations have been requested somewhat recently in the Issues tracker as well.

rmcarthur commented 5 years ago

Are you going to fork off of #129? I really liked the approach of that, and might take a stab to fork and finish up based off the review and comments there.

akacarlyann commented 5 years ago

@rmcarthur Yeah I set up templates using lucasvw's api approach with add_layer and remove_layer methods. I'm actually nearly done, with the exception of legend handling. There's not a lot of real estate, so we'll probably have to introduce some sort of minimization of the legends.

rmcarthur commented 5 years ago

I've actually built something very similar and having similar things with the legend. It looks great when not in an iframe though. I look forward to your PR. I can help review and comment if you'd like. I'm very new to this repo, but needed the layering aspect so I hacked something together fairly quickly. Is there anything I can help with?

akacarlyann commented 5 years ago

Oh awesome! I suspect we are members of a growing club of multi-layer mapviz-hackers. I'd really like to get my PR up sometime this week and happy to collaborate @rmcarthur!

ryanbaumann commented 5 years ago

I suspect we are members of a growing club of multi-layer mapviz-hackers.

Yes, it certainly seems so! Thank you @akacarlyann @rmcarthur for leading the charge. Looking forward to a draft PR to review.