##### The Definitive Statistical Analysis of

by Zack Abrams | Deputy Editor, Bwog

# A quick overview of the project

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 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.

• Unique Posters: 70
• Total Reactions: 33,761

### 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.

# 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.

# 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."

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.

### How do we react?

I was interested to see in 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.

## What did we learn?

Well, not a whole lot. Facebook's algorithm ensures that there's not one good time to post; if your meme does well, it's likely to stay on the top of the page for long enough so that everyone can see it. Price doesn't have much to do with success, so save your money. It's interesting how many memes are barely related to Columbia; however, if the poster thought it was funny enough to share, others probably did too. A higher proportion of 'Sad' reactions went to posts about mental health and classes, which is to be expected. And we need to come up with better captions.

Want to talk to me about this project, memes, or literally anything else? Follow me down below, or email zack.abrams@columbia.edu. Thanks for checking out the site!