A site for reassigning RSA stats data into different geographical areas for analysis

A scientific research company in South Africa wanted to expose national survey statistics in a new format. Rather than grouping the household size and other values by municipal area, they wanted to use more natural geographic and settlement based boundaries.

We used a mathematical model devised by our client to import all stats, rework them, and export them to shape files, which our client rendered in their GIS systems.

A primary difficulty was the size of the data-set, which while not on the scale of Google, required some consideration to achieve a non exponential increase in complexity during calculations.