Just to properly attribute my work and explain a little how it works, the data comes from the gdelt-hathitrust Google BigQuery dataset of which I created a locally stored Postgres database containing a decently large subset of the results (to reduce query costs). The mapping itself uses an OpenLayers OSM map with the tiles being provided by stamen.