We made another run at an in-person event in December, with the hopes of being in-person every other month in 2022.

Because it had been a while since we were in-person, emcee Bryan Huber did some audience participation with some real-time data gathering (sadly, said data was not rigorously collected, so no reporting or analysis will be provided here).

Bryan Huber Surveys the Attendees

We did a light bit of actual content while we munched on pizza and sipped on mulled wine, with Tim Wilson presenting on some of the potential pitfalls of working with time-series data, and how the technique of “first differences” could be used to check that two metrics that appear to be correlated are not simply both trending over time.

Tim Wilson: Possibly a Staged Photo Prior to the Event

Tim explained that first difference is simply looking at the change in a metric from one period to another: if two metrics are correlated, then they should both move proportionally over time.

First Difference Calculates the Change from Period to Period

Correlation of Raw Values vs. Correlation of First Differences

The complete deck is available here: 2021-12 Columbus Web Analytics Wednesday – First Differences

Perhaps the evening would best be summarized with a poem:

There was laughter
But ‘nary a tear.
And pizza eaten
Washed down with beer.

There was mulled wine;
It was the season!
And “time”-ly content
For R-squared reasons!