← Zack Abrams

Data journalism · Bwog · January 2018

The Definitive Statistical Analysis of Columbia Buy/Sell Memes

By Zack Abrams, then Deputy Editor of Bwog · Read the original on Bwog

Columbia Buy/Sell Memes is a private Facebook group where members of the Columbia community, along with other students, post relatable memes about being a student at Columbia or Barnard.

Engagement

Engagement data

On January 25th, I collected data from the last 100 posts in the group that were at least 24 hours old. All of the posts were from December 2017 or January 2018.

70Unique posters
33,761Total reactions

When do people post?

When collecting the data, I noticed that very few students posted memes during the AM hours. However, the results are even during the daytime and the nighttime.

The group Fact-check me: the data set

Analysis

When is the best time to post?

Whenever I post a meme, I always wonder if there's an optimal time to post according to Facebook's algorithm or my classmates' activity. I plotted the total reactions vs. the time posted, but I found no statistical correlation.

How much should I charge?

One of the benefits of structuring the page as a marketplace is the ability to change the price of your meme. I plotted the price of memes on a logarithmic scale, but again did not find any statistical correlation.

Topics

Here's what we're talking about

Topic distribution

Many of the memes take aim at common targets: Econ majors, Canada Goose owners, and campus mental health, to name a few. Here's the general breakdown of categories in December and January.

Many memes start with familiar phrases like "price is" in the title or the caption. Here's the percent of posts that had a sentence start with "price" or "when."

How much about Barnumbia?

Many students simply repost memes from other pages, failing to customize their memes to the Barnumbia community. Others only partially customize their meme by making a reference in the caption.

The question is, how much does this matter to the users of the page? Surprisingly, there's little correlation. Here's a chart comparing the average reactions for each post type.

Reactions

How do we react?

I was interested to see the correlation between certain topics and the proportion of reactions, like if we react 'sad' more on memes about mental health than about careers. See for yourself below.

Originally built in 2018 and rebuilt in 2026 without its old template. The data, charts and word cloud are unchanged.