Test-driven development (TDD) uses agile and lean approaches and test-first practice instead of testing near the end of a development cycle. In this post we will use a simplified example of association rules in retail industry to illustrate TDD in data analysis.

In this post, we'll load a large GeoJSON file into MongoDB by ijson (a Python package). Then we'll answer two questions using MongoDB aggregation pipelines; SQL queries are also given for comparisions. We'll wrap up with visualizing querying results on a map.