I can use search API to specify geolocation with radius and will receive tweets from that location. If I have a larger area, there will be more tweets available than fit into the response. What would be the best way to get a week of tweets from certain region?
My ideas would be:
- divide the area into very small areas and call the search for every small area. Let’s say the initial area is 10x10km, then I will divide it into 100x100 meter areas. This leads to 10,000 queries which is way over the limit. Besides, some of those areas might not have tweets at all (or only have 10 tweets). So, it might be wise to merge some smaller areas etc.
- take the whole area (10x10km) and use max_id to loop back in time. That way I don’t have to take care of the problem dividing my initial program into too small subproblems.
Is there any difference in the outcome? For example, if I take large area and use max_id to loop back in time, may-be some tweets will be missing compared to searching tweets by small area?
If you have any other ideas, how it would make sense to get public tweets from certain region, please let me know.
(btw, previewing my question, the subject and the category selection will become lost)