You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
While reviewing the sample output (thank you for that example!) I noticed that I am unable to reproduce your pytrends results for historical keyword searches.
Namely, the overlapping chart for the "suncream" keyword produces the following output and clearly demonstrates that the search volumes I extract for the (2018-01-01 00:00:00 and 2018-05-15 00:00:00) as well as the (2018-05-16 00:00:00 and 2019-02-09 00:00:00) time periods differs from the example you provided.
Importantly those time periods show extreme outliers in search volumes which results the following long-term trend after scaling:
Because of these outliers, the longtrend.build() method returns the following error: ValueError: unable to scale: to_scale has range 0 in overlap with scale_by; this may be because of an extreme spike in trend data
Do you have any suggestions on how to fix this error or possible explanations as to why the trends are differing so much?
Any and all advice would be greatly appreciated! Thanks in advance!
The text was updated successfully, but these errors were encountered:
Hi @cas2247 , thanks again for the issue. Please can you let me have the exact code you're using? When I plug in those dates, I get the below. I'm wondering if it could be a geo issue, with you being in the US presumably (where 'suncream' is not a well used term), and me being in the UK. If so I'll look into why that is, but if I could have your code that would be a great help. Here are my results:
While reviewing the sample output (thank you for that example!) I noticed that I am unable to reproduce your pytrends results for historical keyword searches.
Namely, the overlapping chart for the "suncream" keyword produces the following output and clearly demonstrates that the search volumes I extract for the (2018-01-01 00:00:00 and 2018-05-15 00:00:00) as well as the (2018-05-16 00:00:00 and 2019-02-09 00:00:00) time periods differs from the example you provided.
Importantly those time periods show extreme outliers in search volumes which results the following long-term trend after scaling:
Because of these outliers, the
longtrend.build()
method returns the following error:ValueError: unable to scale: to_scale has range 0 in overlap with scale_by; this may be because of an extreme spike in trend data
Do you have any suggestions on how to fix this error or possible explanations as to why the trends are differing so much?
Any and all advice would be greatly appreciated! Thanks in advance!
The text was updated successfully, but these errors were encountered: