Geospatial anomaly filtering of geolocation data streams

A method, comprising:

• receiving at a data refiner geospatial data from a third-party data broker;
• wherein the data refiner includes a computer;
• wherein the geospatial data includes multiple data pings;
• wherein each data ping includes a location for a user device;
• determining the data pings for a cell in the geospatial data satisfies a threshold with the data refiner;
• wherein the cell represents a geographic area within the geospatial data;
• analyzing the data pings in a subcell of the cell with the data refiner in response to the determining the data pings for the cell in the geospatial data satisfies the threshold;
• wherein the subcell is located within the cell;
• wherein the subcell has a smaller area than the cell;
• calculating a density threshold condition for the subcell with the data refiner;
• wherein the density threshold condition represents a realistic number of the data pings observed within the subcell;
• wherein the density threshold condition is a maximum ping count for the subcell;
• counting a ping count of the data pings in the subcell with the data refiner;
• determining the ping count in the subcell exceeds the maximum ping count for the subcell with the data refiner;
• removing the data pings in the subcell from the geospatial data to have the ping count in the subcell to be at most the maximum ping count with the data refiner;
• and
• wherein the removing the data pings in the subcell from the geospatial data occurs in response to the determining the ping count in the subcell exceeds the maximum ping count.