One way is in response to what I call "demand," or "traffic." That is, people ask more questions about the weather than about education, so "weather" gets a higher priority in data gathering.
Another way is what I call "mandate" (by department). That is, the health department is tasked with gathering data about various diseases, the education department will have a comprehensive database about all educational institutions, and each department will gather all relevant data that falls under its mandate, etc. That is what I call a "Dewey decimal" approach to data gathering, where pieces of data are gathered according to "maps," each piece of data is considered to have comparable importance to others, and the various departments all have similar budgets.
Which of these models more accurately describes the data prioritization process? Or could it be a "blend" of the two, or even a "third" process that I haven't covered.