A plug-in for glue-jupyter for the loading and analysis of geospatial data. This plug-in uses geopandas to load almost any vector-based spatial data format (ESRI shapefile, GeoJSON files and more) and uses ipyleaflet to provide an interactive map that is coupled to other displays of your data so that selections propagate using the brushing and linking paradigm of the glue visualization library.
You can use glue-map and glue-jupyter for analysis within a Jupyter Lab session and you can transform your Jupyter notebook into a standalone website using voila and voila-gridstack.
- Read almost any vector-based spatial data format (shapefile, GeoJSON, etc) using the geopandas library
- View and interact with your data in a variety of ways: map, 2D scatter, 1D histogram
- Overplot point-like tabular data (e.g. from a CSV) on a map alongside other spatial data
- Create subsets from the map viewer and see how those subsets propagate to other viewers, or see how subsets in other views translate to spatial regions on the map.
- Export your customized Map Viewer to a standalone leaflet.js script for easy embedding in static webpages.
- Serve a complicated Jupyter Lab session as a standalone website using voila and voila-gridstack.
Because several of the underlying packages require very specific versions to work together, the easiest way to get a fully functional environment that includes voila and voila-gridstack for layout is to use conda/mamba to install this plug-in into a new virtual environment.
First, grab the repository code:
git clone https://github.com/jfoster17/glue-map.git
Then create a new virtual environment using conda or mamba:
conda env create -c conda-forge -f environment.yml
Activate the new environment:
conda activate glue-map-with-voila-gridstack
Alternatively, if you already have a working environment for glue-jupyter you can simply do:
pip install git+https://github.com/jfoster17/glue-map.git
A sample notebook is included in notebooks/glue-map-demo.ipynb
This notebook is configured for use with voila-gridstack to enable better layout of multiple viewers. With the recommended environment, you can show the gridstack layout from within jupyter lab as shown below.
Alternatively you can invoke directly from the command line as:
voila glue-map-demo.ipynb --template=gridstack
To edit the existing layout, including adding different viewers or changing their position, user the button highlighted in red within jupyter-lab to change the layout:
Another example:
Since we cannot easily get error messages from glue-jupyter when a plug-in fails to load, do this:
python -c 'from glue_map import setup; setup()' to verify that this plug-in can load.