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
In the app, we currently have access logs which display the search queries and parameters: essentially date/time, search terms, language, category, page.
We don't have any information about sessions, IPs or in general who queried what: the data is completely anonymous. We want to keep it that way.
It could be useful to feed this (again, anonymous) data into an analytics platform, so that we can have a look at the most queried terms. Using this data, we would be able to focus on popular search terms when tuning our app for best relevance: e.g. if jpa is queried frequently, we better make sure it returns good results (BTW it doesn't right now -_-).
The text was updated successfully, but these errors were encountered:
@maxandersen do you have any preference regarding the analytics platform? Maybe we could reuse the platform we use for build analytics?
Technically OpenSearch is able to provide analytics, so we could just index access logs there. We'd probably be able to handle term analysis/normalization better that way. But I'm not sure it would be practical, since we'd end up with e.g. two separate dashboards on two separate UIs for build analytics and search analytics.
EDIT: And there's also the fact that the current OpenSearch indexes are not really important (we can rebuild them), whereas this new data would need to be preserved/backed up/whatever. Seems like extra work.
In the app, we currently have access logs which display the search queries and parameters: essentially date/time, search terms, language, category, page.
We don't have any information about sessions, IPs or in general who queried what: the data is completely anonymous. We want to keep it that way.
It could be useful to feed this (again, anonymous) data into an analytics platform, so that we can have a look at the most queried terms. Using this data, we would be able to focus on popular search terms when tuning our app for best relevance: e.g. if
jpa
is queried frequently, we better make sure it returns good results (BTW it doesn't right now -_-).The text was updated successfully, but these errors were encountered: