The summary statistics for top and rising questions max out at 25 terms, even when the raw data records have well over this many terms that need to be summarized. At the time of this posting, the sample datasets being constructed from daily cronjob executions of GTrends data pulls yields 4 "query" or "terms" datasets: 2 from rising and 2 for top at 2 distinct dates. These 4 files have (in no particular order) 76, 60, 77, and 58 distinct terms in their raw_data_records sheet that should be represented in the summary stats dataset.
Perhaps this has to do with a critical mass of samples? There are terms in the sumstats sheet with calculations for n(observations)=4 though, so I'm not sure. Check to see if there are associations between either number of observations or how early in the raw data records the term shows up and what shows up in the summary statistics. More investigation is needed.
The summary statistics for top and rising questions max out at 25 terms, even when the raw data records have well over this many terms that need to be summarized. At the time of this posting, the sample datasets being constructed from daily cronjob executions of GTrends data pulls yields 4 "query" or "terms" datasets: 2 from
risingand 2 fortopat 2 distinct dates. These 4 files have (in no particular order) 76, 60, 77, and 58 distinct terms in theirraw_data_recordssheet that should be represented in the summary stats dataset.Perhaps this has to do with a critical mass of samples? There are terms in the sumstats sheet with calculations for
n(observations)=4though, so I'm not sure. Check to see if there are associations between either number of observations or how early in the raw data records the term shows up and what shows up in the summary statistics. More investigation is needed.