Algorithms

So, the question of creating appropriate search strategies for effectively utilizing the GIS and survey data comes down to envisioning how the data can be sifted to find very specific events or characteristics.

This is an interesting academic challenge because it requires me to explore the limits of my analytical and logical thinking while missing some rather significant chunks of content. For example, I can write a calculation syntax that allows the database to report out the farmers who have timber of a certain type, calculate how close that timber is, on average, to the closest road, and use that calculation in a rating system to identify farmers who may be the best match with Saaraketha's goals and objectives. What I can't do, however, is know for certain what an optimal distance is for timber from the road.

Of course, there's a theoretical optimal distance from the road of zero meters (imagine a farmer who has timber growing right on the side of his road), but that may not be the practical optimal distance. Also, what is a "likely" result for this mean distance for it to be meaningful. Without that bit of content, I can't really set up a very good ranking system because I can't say whether 10 meters should receive a maximum score or whether 25 meters should.

This becomes even more important when trying to rank someone on a variable associated with the value of the timber that they have. In order to weight the value appropriately, I need to transform it from a raw number (e.g. 36,000 Rupies worth of timber available) to a weighted score (e.g. 36 our of 40 points available). It is pretty straight-forward to calculate the worth of someone's available timber. After all, it's simply the number of trees of that sort times the percentage of trees that are harvestable times the value of each harvestable tree. Where my knowledge kind of falls off the cliff is in trying to see what an "optimal" timber worth score would be for this setting. Is 36,000 Rupies a "normal" score or a "high" score? Is it really low? Would it be more likely that a farmer would have 500,000 Rupies worth of timber? Without knowing the answers to these kinds of questions, I don't know if a 36,000 score should receive high points on a weighted scale or low points.

As I said, this is an interesting intellectual exercise, but it's tremendously difficult to pull off without that kind of foundational knowledge.