When people break up, according to Facebook updates

For his book The Visual Miscellaneum, David McCandless, along with Lee Byron, had a look at breakups on Facebook, according to status updates. They looked for the phrase “we broke up because” in status updates, and then graphed the frequencies over time. Why they couldn’t just look at updates to relationship status, I’m not sure.

Notice the peak leading up to the holiday season and spring cleaning. Then there’s the people who think it’s a funny April Fool’s joke to say the broke up with their significant other.

Finally, there’s the highlight of Mondays, which you might lead you to believe that people like to call it quits during the beginning of the week. My hunch though is that it happens towards the end of the week, people use the weekend to be sad, and then talk about it on Facebook.

[Information is Beautiful | Thanks, Elise]


  • My hunch was that the weekend was used to get laid one last time and that’s why the break -up took place on Monday.

  • My thoughts on this are a) aren’t Mondays just the most likely time you’d mention on facebook that you’d broken up with someone at the weekend? And b) On another site they said they’d scraped Facebook for various phrases, not just “we broke up because”. One of which was “break up” – as in “we’re spending spring break up at her parents” or similar. Might help to explain the size of that spike.

    Also, I too wonder why they didn’t just search for the relationship status updates. Would seem to be a fairly standard phrase that’s easy to grab.

  • I think they didn’t use the relationship status because it’s much more complicated to know the exact time of a relationship status change than a status update using Facebook public APIs.

    Using these APIs, you can easily grab all status updates, and the date they were posted is given to you. But you can only know the relationship status, you don’t have any information as when it changed for the last time. Of you want to have this information, you’ll have to get the information every day and check up for changes manually.

  • The “monday phenomenon” seems to occur in April and May, but not in the summer. My hypothesis is that younger people account for a larger percentage of break ups as well as students are effected by the summer schedule. I’d love to see some age frequencies.

  • I wonder if the holiday peak is due to people not wanting to bring a soon to be ex home for the holidays or just a reflection of people thinking more about their lives and happiness before the end of the year. Probably both I guess.

    What are the odds of doing multiple graphs like this but for different age groups?

  • Any ideas as to what is happening in mid-March to diminish the amount of breakups? The first blossoms of Spring? (If so, then the rising amount of break-ups from Christmas to mid-March could well be an indication of Seasonal Affective Disorder.)

  • I would like to see the graph normalised for total status updates per day. I for one don’t change my facebook status on Christmas day but often do it when I arrive at work on Monday morning.

  • The Monday phenomenon makes perfect sense – it is due to what happened over the weekend, typically the time that couples get together – an infidelity, bad date, a bad habit that was finally the last straw, or just a realization of weariness and monotony after getting together. The weekends are also when singles that have a “night with the girls/boys” might run into a new Mr/Miss Right and then clear the path on Monday.

  • I like the “Spring Clean” and “Too Cruel” remarks – how true yet hilarious.