Hi,
I’ve been watching for a few times Sara Bird’s talk on choropleth maps with bokeh here Sarah Bird - Interactive data for the web - Bokeh for web developers - PyCon 2015 - YouTube
and thanks to that video I got a basic map working with polygons (coming from a GeoJSON projected to Web Mercator EPSG:3857),
filled with colors depending on values from a pandas dataframe, using STAMEN_TERRAIN as tiles provider.
I was wondering how could this choropleth map be refreshed with filters coming from a selector without a server interaction?
I know in Sarah’s video it is mentioned something about “actions”, and I digged a bit in the source code checking things regarding CustomJS,
but I am not sure how to procede.
It looks I should somehow inject some custom javascript into the python API, but I struggle
to find hooks in the examples and I am not really sure where should I look in the docs.
Ideally I would like to apply that filter from a selector to a pandas dataframe like a “groupby” where the
filtered records have all the same structure so then the choropleth map finds the fields where expected.
The selector is a year quarter e.g. for today July 2017 is “2017-Q2” as a value for the cell in the rows to filter.
I am working with jupyter notebooks at the moment, but I would like to run a python script
to generate static HTML files once this filtering thing works without the need of a backend.
This is an extract of the code:
source_data = ColumnDataSource(data=merged_df)
fig.patches(‘xs’, ‘ys’, source=source_data,
fill_color={‘field’: ‘median’, ‘transform’: color_mapper},
fill_alpha=0.5, line_color=“black”, line_width=0.5)
This is the head of the dataframe:
year_quarter,my_polygon_id_name,count,min,max,mean,median,xs,ys
2017-Q2,MY_POLYGON_NAME_VALUE,108.0,531.67,80000.0,2266.992315,1527.27,(the longitude coords),(the latitude coords)
Thank you
Alessio