Tableau Data Transformation Detail:
While Google Analytics feeds the web traffic data to Ohio State’s Tableau dashboards, there are several important differences between the data as reported by Google Analytics tools (GA interface and Looker Studio) and the data in the Tableau dashboards. They come from the data transformations that are performed in our data warehouse before the data is summarized for reporting.
- Bot traffic. Weekly traffic spikes from automated crawlers, like SiteImprove, and other bots are removed from the Tableau dashboards. This is by far the largest difference in the data.
- No pageview sessions are excluded in the Tableau session counts but are counted in Google Analytics. Other metrics, such as engagement time, engagement session indicators, or outbound link clicks that happen during these sessions are generally included in the Tableau reports.
- Attribution - Channel definitions differences.
- Internal Channel. In Tableau, Internal channel is traffic from internal sites and apps, such as Zoom, Teams, Sharepoint, and the Wexner Medical center. In Google Analytics, this traffic typically appears under Referral channel.
- AI referrals. Traffic from ChatGPT, Perplexity, and other AI sources is reflected in Tableau.
- Video channels. In Google Analytics, YouTube and TikTok traffic appears under Organic and Paid Video, while in Tableau, we roll it into Social traffic.
- ‘Other’ channel. The definition of ‘Other’ channel in Tableau is wider than that of Google Analytics’ Unassigned channel and covers rarely used channels, such as SMS and Shopping traffic.
- Non-pageview channel. This channel is used to identify events like Outbound link clicks when they happen during a session with no pageviews. This typically happens in an interrupted session, when a page remains open in the browser for over 30 minutes.
- Unique Pageviews. Tableau reporting includes Unique Pageviews metric, which is currently not available in GA. Unique Pageviews are a count of unique sessions on a page. They should match session counts for individual page URLs in Google Analytics. However, unique pageviews on a hostname or site level are higher than sessions because a pageview is counted on every distinct URL. In Tableau, session count metric is not available on a page URL level (use unique pageviews), but is available on a landing page level.
- Go link redirects. Go Link redirects originate from a custom event in Google Analytics and are available in Tableau. However, Google Analytics UI has no built-in reports for custom events like Go link redirects.
- Users. Currently, Tableau reports do not include user counts because it is impossible to calculate unique users from the summarized tables. We are looking at other solutions that would allow reporting unique user counts.
- Engagement rate over 100%. On occasion, a session without any pageviews is an engaged session. In that case, the session is not counted in total sessions but counted as an engaged session. For example, when many visitors to a page let the session lapse and re-engage later, then the engagement rate may climb over 100%.
- Page URL. In Tableau, all page URLs are cleaned for query strings and utm parameters. They are also standardized to remove ‘/’ at the end of the URLs.
- ‘Other’ page URLs and hostnames. Occasionally, the event data contains broken, numeric, or localhost URLs and hostnames. In Tableau, they are cleaned and summarized under “Other” category.
- Session source. Session source comes from a parameter under umt_source or is designated by Google when processing the data. In Tableau, many common traffic source values are cleansed and summarized. For example, m.facebook.com is replaced with Facebook.
- Not set. Google Analytics shows ‘Not set’ when the value is not available in the breakdown. For example, ‘Not set’ is the value assigned to the landing page URL in sessions with no pageviews. Tableau data rarely includes ‘Not set’ values but may show Null where the value is not available.