Peak Break-Up Times On Facebook

Wednesday, November 3rd, 2010

An image myself and Lee Bryon created in 2008 for The Visual Miscellaneum using facebook status updates

(Somebody redrew this graph and now it’s going round like wildfire but without credits – gnash! So I thought I’d officially put it up)

Christmas too cruel. Hahah! Still makes me chuckle.

Peak Breakup Times according to Facebook

Books and Store

Our Beautiful Books - Information is Beautiful Information is Beautiful Store

Show Comments ( )

  • gert

    sorry to hear that somebody is mis-using your nice work. I really appreciate what you are doing, keep it up!

    • http://www.seo-saijogeorge.com.au Saijo George

      That sucks , try contacting the guy to see if they will add it now

  • http://www.Sumerian.com Blair robertson

    This would have to be normalised against the distinct count of users each day. I’d expect less login activity on (for example) Christmas day and therefore less status updates.

    • http://dejanseo.com.au/ Dan Petrovic

      Very good point. I didn’t think about that.

      • Ron

        The same could be carried over for Monday’s. Facebook usage down over weekends and people playing catch up on Monday’s. With advent of smart phones I would argue these non usage false data ‘peaks’ would be flatter.

  • Mark Beaches

    also got a mention on Nov 3 episode (1343) of CNET’s Buzz out Loud.
    http://www.cnet.com/buzz-out-loud-podcast/
    at about 24:27

  • http://www.velvetslist.com Velvet Ballard

    I wonder how much this has changed in the last two years with more of the general population joining facebook and it having a smaller ratio of college age users. Do 35-year olds follow this pattern? Forty-three year olds?

    Thanks for the data. Clearly there is a fear of unrequited gift giving!

  • Wendy

    I think this chart could be just for my stepson’s babymamma’s facebook status.

  • Johnee_Thunder

    I’d love to see an overlay of when people “get together”. I’m curious to see what sort of overlap there is – does the data mean people like to be single over the summer, or does it mean they like to find someone new before the summer begins?

  • Jeff Larson

    Any way to create a similar “hook up” graph to see if there’s a correlation?

    • http://seojoey.com/ Mike

      Would be interesting to see the correlation. Social data like this could prove very useful.

  • Steve G

    Now that Facebook has a ton of European users who don’t celebrate valentine’s day, what would happen if you mined that data for breakups, and looked for Valentine’s day? I’m amazed that holidays have such a huge influence on our romantic lives, but then I’m reminded that my January breakup with my ex was in part driven by seeing his mother over Thanksgiving.

    • Golo

      Steve G – not sure where you’re getting your information from, but I’m British and Valentine’s day is certainly ‘celebrated’ here, in that it’s customary for lovers/spouses to give each other flowers and cards, go out to dinner etc. Much the same applies in France and Italy too, for instance.

  • http://www.niksnexus.net/weblog nik9

    Enjoying your visualizations. So I have to ask, how do you gather that kind of info to visualize!? Do you actually have to log each status update in this case!? (wonder if the trend continues over time)

  • http://www.niksnexus.net/weblog nik9

    Ah – found same Q and A on your FB post – Facebook Lexicon now extinct – bummer. But I remain curious. Did you see johan bollan’s / tweets as mood indicators for example – not visualized, but that question remains – how to begin to gather that kind of social network data!??

  • http://twitter.com/helencurry Helen

    I love that peak before Christmas – clearly people wondering “shall I buy him/her a present?”. Seems to be a critical measure :)

  • Colin Sachs

    Lovely infograph of the data set… but what does it really tell us? That 10,000 Facebook users, within a single year, had more status-changes which indicated a break-up during two distinct periods of that year… that’s about it. This can not be generalized to any other larger population.

    For all anyone knows, without additional demographic information and without running this over several years, this could just be a pattern for a single year.

    A more accurate title would be: Peak Break-up Times for a Set of Facebook Users Over a Year.

    • ChrisChobez

      Nice trending as well…and I could see this making sense– breaking-up before Xmas so you don’t have to spend more $$ on holidays and write before Spring Break, when people want a change or (for party-goers) liberation from their partner before they go party for Spring Break..

      but I’m desiring to know if this data accounts for someone breaking-up with the same person, again? And for that matter, is 10,000 random people? What about people who say they are dating someone, and they really are not- (like, they’re just friends and having fun on facebook)? And are facebook status updates different than, someone just changing their information from “single” to “in a relationship” in their profile (vs. actually giving a status update in the user’s own words)?

      Anyhow, fun food for thought, thanks for sharing!

    • http://deanhewson.tumblr.com Dean Hewson

      Seems pretty clear that it’s facebook specific when the second line of the title is, ‘according to facebook status updates’ . Don’t see the part where he makes assumptions about it’s relevance to the entire population either…

      A list of asterixes (asterii? lol) with justifications/explanations turns a visualisation that people might actually look at into just a few more 1′s and 0′s in a digital ocean. Since mis-interpreting that graph isn’t dangerous, leaving the qualifications off doesn’t hurt anyone and makes it more likely that the authors’ unique insights will be spread.

      In short: Chill, pedant.

    • Chris

      I’m pretty sure 10,000 would be considered a statistically relevant sample size, and given we are talking about breakups, which occur amongst people who are dating and therefore most typically 18-30, I would say this is very accurate given Facebook is made up of young people (despite the growing percentage of ‘adults’ on there).

  • Peter Fabian

    Still very interesting infographic. Would be interesting to see how does it match with getting together statistic.

    Colin, callling it Peak Break-Up Times On Facebook, or Peak Break-up Times for a Set of Facebook Users Over a Year is really the same. If you wanted to be more specific, you would have to introduce friend statistics of the person. For example, spring break as indicated is strictly USA phenomenon.

  • Michelle Marien

    Well, if that was 2008, the end of my marriage is nicely documented on that second spike in July. How creepy.

  • http://pootling.net Thom

    You say that Christmas break-ups are rarer because they’re ‘too cruel’, but I suspect that it’s down to:

    1. People breaking up on Christmas day at pretty much the normal rate, but not having their normal access to computers, being in someone else’s house (e.g. with family)
    2. People not breaking on Christmas day, but only because they don’t actually happen to be with their partners as they would be on most other days of the year.

    I mean, clearly I don’t actually have any evidence for this, but I think it’s a passable theory.

  • Mike

    Colin Sachs does not understand statistics.

    • Colin Sachs

      Actually, yes I do understand statistics. I’m an analyst (currently) for a living doing BI/Dashboard tracking QI/PI activities… and also a bit overly-sensitive to how data can be presented to misinform.

  • Alice

    boohoo – it’s true. And now my heartbreak can be reduced to a mere statistic. LOL My marriage ended at the peak time in March, though I didn’t broadcast it in my status updates and nor did he.

    I think standard statistics tend to show divorce rates spike near the end of holidays when people who don’t usually spend that much time together suddenly do. That’s not shown up in this sample though. Interesting and fun to look at anyway.

  • Anne Brown

    Do you exult in the breakups of others? Yuck.

  • http://igmorrison.com igmorrison

    I love how Christmas is too cruel, but two weeks before is just fine. Genius!

  • Bill

    Dear Colin

    I don’t know if it is inaccurate or not,BUT did it ever occure to you that if you do break up in one of those two peaks,you can find comfort in this “false” infograph?That’s all it matters anyway because you just broke up,messed up your life….well at least changed your status :)

    *Nah the people who use facebook is mostly 15-40,spent a serious amount of time with their PC,refresh their status,care about their profile,etc so if you are indeed into fb and choose to have a relationship status,you can say it describes you :)

    Anyway it can’t be bad to becareful not to break-up with your boyfriend/girlfriend.

  • http://www.explorechi.blogspot.com Adam

    I think it may be a good idea to match these statistics with the incidence of getting into a relationship. The incidence ratios between the two can shed a lot more light. Some of what we see may just be an artifact of a peak or a lull in relationships.

    One thing that struck me was that the breakup rate around Valentine’s Day is still higher than the base rate. It could be that many are hooking up prior to, in an act of desperation, the failed ones then also ending soon afterward.

  • Jimmy Asher

    I love this graph…. One thing that would be interesting is to correlate the data set with the volume of posts, based upon the fact that I recently heard that the majority of Facebook users update their site a work (explaining the Monday pattern).

  • JB

    The data is also limited to FB users with no privacy settings, so for all those who do set their privacy settings, no data …

  • Jennifer

    It would be interesting to see an overlay of when people start relationships as well in order to see the entire cycle.

  • Matt

    Wow, Colin, you must be the reason that demographers and statisticians get invited to all the great parties. What does your comment really tell us? You’re a ass that is awake at 3:13 in the morning critiquing the statistical responsibility of a FACEBOOK GRAPH. And despite my limited data, I am willing to be that you have been as ass over several years.

  • http://www.theunqualifiedeconomist.com The Unqualified Econ
  • http://coolinfographics.com Randy Krum

    Hey David,

    Your findings (not the visual) were featured on the NPR show “Wait Wait…Don’t Tell Me!” on the November 6th episode. You can find it at the 35:40 minute mark. It’s also available as a podcast through iTunes.

  • http://www.articely.blogspot.com fahmi

    Still very interesting infographic. Would be interesting to see how does it match with getting together statistic.

  • http://dotlearnt.com Anna

    When are peak engagement times? Or, when are the peak times people enter relationships? I wonder if they align to the times above!

  • http://www.homelist.org HomeList

    Sorry to hear you’re getting your content hijacked, but it’s a really interesting data nonetheless!

  • Kerrin Naude

    Great idea, and probably demographically significant. Only concern is that I know a whole bunch of friends who never ever broadcast whether or not they’re in a relationship.

  • Matt

    This is great! A Fourier transform of this might be interesting to see what periodic features are present.

  • Seamus

    So, am I the only one that doesn’t think typing “break up, broke up, breaking up, broken up” into Facebook Lexicon counts as thought-provoking?

  • http://www.enditquick.com Turk5555

    Imagine having a partner with their anniversary in December, plus Christmas and New Years celebration. Then, the same partner has their birthday in February just before Valentine’s Day. Break-up, quick is my solution. This chart is just too coincidental, but true.

  • -

    spring is the [blank]-year itch effect?
    early winter, hmmm

    any difference depending on latitude (or at least hemispheres)?

  • Shanna

    Where are you share buttons? Not top and center for ultimate convenience in sharability…. anyways, I would also like to see the lifespan of relationships vs the actual average lifespan of relationships, and broken apart by age group perhaps? All-in-all, very nice!

  • http://www.primbs.de Dirk

    Any chance that you share with me how you got that facebook data?

    Cheers,
    Dirk

  • Brande Barrett

    I was married on April Fools Day!

  • http://www.abiolatv.com Love Blogger

    I am referencing you and, of course, giving full credit where it’s due. Well done, and happy holidays!

  • Gretta

    What is going on in March? My theory is that March is to relationship what the period of time between winter break and spring break is to high school kids. It gets so boring! You’ve got Valentine’s day and then what? Nothing until spring fever hits again. Luckily my anniversary is in March, good time to cash in on sweet romance.

  • Dave

    LOl, I guess some if those April fools gags backfired big time :D

  • http://www.toysie.com.au/ Sally

    How is information like this obtained? Very interesting

  • arturo

    yeah my babe hates me . im so lonly and willing

  • http://exbackreport.com Devin

    Can you comment any more on how you got the data? I’d LOVE to see the inverse of this graphic: when are the peak times for starting new relationships?

  • http://soundlessvalley.wordpress.com Austin

    I got a kick out of the spike in break ups during spring break and two weeks before Christmas. It’s a safe move considering you won’t see each other for a while, haha. It could also be pretty jank move if you bought your “significant other” something expensive for Christmas. I’m actually shocked summer break ups weren’t higher than April Fool’s Day. Getting dumped on April Fool’s would blow lol.