Talk:Product Analytics/Glossary

For consideration: Sampling

@MWang (WMF), @Phuedx (WMF), and I want to add sampling and need to consider the multiple contexts where it comes up. Maybe have a single entry (section) for it and then specify the different definitions like:

Analytics sampling

Picking some units such as users or sessions for data collection

Data sampling

Choosing some part of a dataset at random for usage in data analysis

Experiment enrolment sampling (bucketing)

Picking some units to be enrolled in an experiment and assigning them to groups

Oversampling

Using a higher sampling rate than the default for a subset of units

– MPopov (WMF) (talk) 18:24, 14 October 2024 (UTC)

I've created a glossary page for Metrics Platform that references the definitions on this page and adds a few terms specific to Metrics Platform. I've included Analytics sampling and experiment enrolment sampling (bucketing) since users of Metrics Platform need to know these terms, but I support using this page as the canonical source for analytics-related terms. Once new terms are added to this page, feel free to remove them from wikitech:Metrics Platform/Glossary. APaskulin (WMF) (talk) 23:35, 28 October 2024 (UTC)