Hello,
My project involves collecting count data satisfying query over several months. For disclaimer purposes I am using the academic research license. One thing I have noticed is that if I perform a request to get counts (even for a time window of 2 days), I will often get slightly different results. This is particularly noticeable if the scale of the data is large, like in the thousands. I’ve been trying to find an explanation for this, but I have been unsuccessful in doing so.
Because of the endpoint paginating at 31 days per response at most, I’ve automated my code to submit requests one month at a time. Sometimes, I am returned data missing certain days, or an empty dataset altogether. While I have found workarounds for this (namely redoing the request if days are missing), I am curious as to why this occurs.
Fortunately, for my statistical modeling purposes, important aspects of the data such as autocorrelation remain virtually identical across runs, even if the numbers are slightly different, so my analysis is unchanged. So this issue does not necessarily need to be “solved,” as I am just wondering why these things may occur.
1 Like
