Attached is a detailed, 13-page spec (the screen shots account for most of the pages), of this web data extraction/scraping project. You will be building a routine/program to extract the wagering and odds data (i.e., the information on the "tote board") that are created in the minutes leading up to each race that is run at most of the U.S. greyhound race tracks each day.
The extracted/scraped data can be saved in whatver standardized format you prefer (e.g., DBF, CSV, XML), but you will need to save the records in a single file that is capable of being read by another application at the same time you are writing/appending to it. The data elements for each record are specified in the attached PDF.
Check out the spec and feel free to post questions.