We sent 10,000 messages. We gave out 10,000 likes. Tommy talked about food at least 50 different times. Here's a breakdown of our house in a groupchat.

Day by day posting volume. Every notable spike is related to a party.

Use by time of week. Unfortunately, I don't know what people's class schedules were because it would have been cool to overlap those with this heatmap. Monday and Thursday are trash day, and pretty much any post made on a Friday was party related.

Our relationships based on who gave who likes. The size of each person's arc represents how many likes they received overall. The ribbons that connect the arcs depict a couple of things. Mouse over Luca's arc, and if you look at the ribbon connecting Josh and Luca, you'll see that it is thicker at Luca's arc than at Josh's. That means Josh gave Luca more likes than Luca gave Josh. The ribbon also depicts this ratio by taking on Josh's color instead of Luca's. You can mouse over the arcs to highlight that person's relationships.

How many times each of us posted, gave out likes, and received likes. This bar chart is with self-likes removed (this mostly only affects Pat and a little bit Josh)


Clustering our vocabulary in September and then again in April. Did living together for a year cause us to borrow each other's language more often? Not really, but the pictures are pretty and there's some increased variation in some clusters. Also, since I generated the graphs separately, the x-y coordinates of cluster centers are random. So that's why Charlie's cluster appears on two opposite ends between graphs. The absolute location of a point does not matter so much as its location relative to other points.