Having spoken with the Product Analytics team, the general consensus is that the aggregations of the raw sessionTick data can be computed in the following:
- As a data scientist or analyst, I am able to query the SessionLength data to display the following distributions
- Average Session Length
- Minimum Session Length
- Maximum Session Length
- Median Session Length
- Distribution by Quantile
- Distribution by Count "bucket"
- Sampling is a requirement to be decided upon before deployment to all wikis, but does not block this work to be done.
Data-QA sign off
- Data in the sessionized table mforns.session_length looks good