Intelligent Weather Forecasting
Intelligent Weather Forecasting with Twitter
Weather forecasting can be important for a number of reasons. At one extreme, prior knowledge of forthcoming weather can save lives (i.e. boats on the sea, fog coming in on a mountain etc). It can also be important for less life threatening situations such as wedding, party events or trips to the beach. Weather forecasting is a difficult task and even when accurately predicted, most weather forecasts from sources such as national TV and radio are quite generic and of little use. What however if we could achieve weather predicting at street level? Impossible, well not if we build on comments concerning the weather from people in the real world. Puzzled? Well, what if we were to 'extract' words from conversations in a mobile application such as twitter where people input keywords such as sun shinning, getting wet, cold, freezing etc.
Twitter is a service for family, friends and colleagues to simply ascertain what each is doing at a particular moment. It is social networking at its best. This project is concerned with extracting words from twitter in an attempt to build the kernal of a system which could be deployed to test the practicality of implementing an accurate fine grained weather prediction framework.
Task
Visit the twitter site. Read the faq. download, and use. Investigate the API. This is the core of what you will work with to develop the prototype.
Home