Stats Processing Plant (Project Dimes)
Stats. Reports. Most people deal with stats in their daily lives. Even folk who never go near a computer.
This system would take in feeds (you could kick off with just XML), and intelligently work out patterns based on the input.
There would be four parts, that I can see.
- The Parsing of the Feeds (the different types of file structures)
- The Rules Interface
- The Interpretation of the Rules / Calculation of Results / Pattern Recognition
- The Reporting of Results (Bling Graphs and wotnot)
There are so many sub-systems you could build around this. Contextually refined for various tasks. But this one would be generic. Meant for the masses.