Wikimedia Product/Data dictionary/session_length_daily

The wmf.session_length_daily table (available on Hive) contains 'pre-aggregated' session length data, and aggregated over a predefined set of dimensions. It is stored in the Parquet columnar file format and partitioned by (year, month, day).

This page describes the data set session_length_daily that is loaded from wmf.session_length_daily on Hive through Presto, which can be accessed via Superset. wmf.session_length_daily on Hive is a derived and aggregated table of the raw event table event.mediawiki_client_session_tick.

Schema

Field name	Data type	Description	Data example	Source schema	Source field
wiki	string	Wiki project that had an active session	ar.wikinews	event.mediawiki_client_session_tick	meta.domain
session_length	int	Time (in minutes) users engaged on a page of that wiki	15	event.mediawiki_client_session_tick	tick
session_count	int	Number of sessions with this session length	100	event.mediawiki_client_session_tick	COUNT(*)
year	int	Year of active session	2021	event.mediawiki_client_session_tick	year
month	int	Month of active session	2	event.mediawiki_client_session_tick	month
day	int	Day of active session	12	event.mediawiki_client_session_tick	day
Following Fields are available in Superset ONLY
date	string	Date of active session	2021-02-12	wmf.session_length_daily	year, month day
bucket_session_length	string	Bucketed length of sessions	0-1m, 1-5m, 5-15m, 15-30m, ..., 12h+	wmf.session_length_daily	session_length
project_family	string	Project family name	wikipedia	wmf.session_length_daily	wiki

Dashboards which use this table

Session Length Dashboard