I Just Tweeted, To Say, I Love You…

In which part of the United States do you think people say “I love you” the most?  I was curious about this, but since I can’t eavesdrop on every conversation (I don’t have any friends in the NSA.. that I know of) I decided to take a crack at answering the question using Twitter data.

Not all Twitter data is public, but there are plenty of public Twitter accounts, and you can look at the data and run your own analysis. (see technical notes below)

Tweet _I Love You_ Map

What did I discover?  First off, the state with the most tweets saying “I love you” (adjusted for population) is… Hawaii!  Aloha!  Who wouldn’t be full of love in the land of sun and sand?

Next up, Maine, and.. Kansas??  I didn’t think Midwesterns were so expressive!

And the state with the least number of tweets, per number of people?  That’d be West Virginia.  Poor West Virginia!  I’ve driven through there and I have to admit all that highway construction would probably make me less likely to feel the love, too.

If you’re wondering who people are declaring their love for, 3% of all the tweets were for Justin Bieber.  He beat out all the guys in One Direction, as well good ‘ole Mom and Dad!  (Moms got more ‘I love you’s than dads did, by the way.)

I’m not going to call this an incredibly scientific study, but it was a lot of fun nonetheless.


Technical notes:  To get my Twitter data, I used Ruby (TwitterStream gem).  You need to have a Twitter developer account.  Once I had my raw data, I ran it through KNIME, a free data analytics tool, and finished up with more Ruby scripts.  The map comes from Google Maps templates (also free).  My Ruby code is available in GitHub (username cpearl42).


Mmmm, Sweaty T-Shirts

question-mark-shirtYou exchange emails with someone, from friendly to flirtatious.  You like the way they look in their photos.  You finally arrange to meet in person.  And… nothing.  No sparks.  No “chemistry”.

But what is chemistry?  Why are you attracted to some people and not others? One piece of this puzzle is something called pheromones, which are chemicals we secrete.

As it turns out, we’re attracted to people who have a different MHC (major histocompatibility complex) than us–that is, whose immune systems are not the same as our own.  This helps ensure any offspring will have a better immune system than one from two people with similar MHCs.

You could do a blood test on a potential mate to check the 6 genes that will tell us this information, or, you could just trust your nose.  Pheromones will do this job for us–if you really like the way someone’s armpits smell, chances are they have a very different MHC than you.

There have been multiple experiments that bear this out.  Subjects wear a t-shirt for a couple of days, which is then placed into a jar or plastic bag, and then smelled by members of the opposite sex.  The people doing the smelling report on how pleasant or unpleasant the smells are, and the ones whose smells are pleasant have different MHCs.

There have even been pheromone parties, in which willing datees bring their own smelly t-shirts to be sniffed!

I say it’s time to bring this to the on-line dating world.  One way would be to supply a blood sample for your dating profile, so you could eliminate people whose MHCs are the same as yours (since you probably won’t like they way they smell anyway).  Another way would be to send in your sweaty t-shirt, which could then be analyzed and quantified.  People could buy a “pheromone tester,” a device which could reproduce the smell of a potential match and you could smell that person in the comfort of your own home.

(As I was doing some reading for this blog post, I found out I’m not the first one to propose using MHC matching in on-line dating profiles.)

On-line dating sites are using algorithms to predict compatibility, but we need something that predicts chemistry as well.  Imagine only meeting up with potential partners who you already know you’ll feel an attraction to.  The technology exists.  Someone just has to build it.

At the Corpus, Corpus Callosum..

Corpus_callosumThe corpus callosum is a wide bundle of neural fibers that connects the left and right hemispheres of the brain.

In 1982, an article published in Science suggested that the corpus callosum is wider in womens’ brains, allowing for better “cross-talk,” which is evidence for womens’ greater ability to multi-task and superior language abilities.

However, a later analysis that looked at 49 prior studies of the brain found that in fact there are in fact no differences in size between the male and female corpus callosum.

The problem is, the original study is what stuck.  Years later, it’s still a commonly-held belief that women have this greater connection between the sides of the brain, and therefore “scientific proof” their communication style is different.  Here’s some recent advice from Dr. Pat Allen, relationship guru:

Number one all women have a bigger corpus callosum (the thick band of nerve fibers that connects the right—feeling brain lobe to the left-thinking lobe) than men.  I recommend the book, Sex On The Brain by Daniel Amen. Since women have a bigger corpus callosum, we can mush everything together. Most right-handed men have a smaller corpus callosums so they’re either thinking or they’re feeling but not at the same time so you have to get them in one lobe or the other. When you talk first, you’re in left lobe he’s in right lobe.  And if you speak first you’re setting up a program where you’re now ‘the leader’ and if you want to be ‘the leader,’ you got a follower or you’ve got a man who’s masculine and who will be bored stupid with your ‘masculine thinking,’ and say “Yeah lady, bye.”

Those poor men, they can’t think and feel at the same time!

So what does this have to do with on-line dating?  Dating sites are always competing to improve their algorithms.  Neuroscience is one more tool available to them.  Here’s a recent study that used MRI to look at speed daters’ brains prior to the speed date to predict who’d they match up with:

When the researchers looked back at the brain scans of volunteers who viewed photographs of people they would later meet, the team found that the behavior of two parts of the brain’s dorsomedial prefrontal cortex — a region near the front of the brain that sits above the eyes — could predict whether viewers would later pursue the people in real life.

There could be a way, in the future, to take advantage of this type of brain information.  But we must be cautious.  Building a dating app based on false science could simply add another wrong path on the already tricky road to matchmaking.

Google Shares the Love, Facebook Shares the Stats

GoogleVDayGoogle joined forces with Ira Glass and This American Life to produce this year’s Valentine’s Day Google Doodle.  When you click on a candy heart, you hear a (true) short love story, accompanied by animations.

Google has a dedicated Google Doodle team, and recently added engineering staff to the team so they could move beyond the static image.

Facebook has been posting relationship-related stories all week from their Data Science team, and today’s post gives some insights on frequency of posts between two people who are getting ready to change their Facebook relationship status from “single” to “in a relationship”:

During the 100 days before the relationship starts, we observe a slow but steady increase in the number of timeline posts shared between the future couple. When the relationship starts (“day 0”), posts begin to decrease. We observe a peak of 1.67 posts per day 12 days before the relationship begins, and a lowest point of 1.53 posts per day 85 days into the relationship.

What exactly is going on twelve days before people “make it official”?  The Facebook Data Team speculates people may be spending more time in person instead.  I’m sure, however, some of these relationships are be long-distance… perhaps people are switching to more texting and talking on the phone?  Or perhaps the communications are becoming more intimate, so they’re not shared publicly.  It could also be that as you know you’re heading to exclusivity, you expend less courtship energy.

Having met my husband in the days before texting and Facebook, I sometimes wonder what would be different about our early days of dating.   Does anyone save their texts or Facebook messages?  Or are these new methods of courting ephemeral, like talking on the phone?

Profile photos: good or bad?

I just finished a fascinating book:  Data, A Love Story: How I Cracked the Online Dating Code to Meet My Match, by Amy Webb.

Webb was 30 and looking for a mate, but found her on-line dating experience disappointing.  She loves data, so she decided to game the system:  first, she created a matrix of qualities she was looking for in a mate, along with a scoring system.  Next, she created 10 fake male profiles on JDate, and checked out the competition.  What were women with successful profiles doing differently?

Amongst other things, their profile photos.  She found the most successful profile photos had several things in common:  the subject is by herself; good lighting; a genuine smile; show some skin.  After a month of research, she took new photos, created a new profile, and set to work.  She started scoring potential matches against her matrix (they had to have a minimum score before she’d go on an in-person date) and after not too long… she found her match.  Who turned into her husband.  Her strategy worked.

Dating site Zoosk (third most popular dating site after Match.com / eHarmony) recently outlined some tips for creating the best profile pictures, based on their data, and say it’s important to take some photos outdoor, and have at least one that is a full body shot.  According to Wired , selfies for women’s profiles are fine, but not for men.  And don’t post photos of yourself with an animal–it’ll reduce your chances 53%.

Sometimes it seems like people ONLY look at profile photos.  Alli Reed decided to try an experiment.  She created a profile of a horrible woman who said she’s really good at “convincing people I’m pregnant lol” and one of the things she could never do without is “keeping america american”.  BUT. For her profile, she used a picture of a friend who is a hot model.

Not surprisingly, she got a thousand replies.  Initially, she figured, again, that the men simply weren’t reading her profile, just going by the photo.  Then she started chatting with them.  And it turns out they didn’t seem to care if she was a horrible person–they still wanted a date.

All this leads me to wonder if we’re doing it all wrong.  Maybe profile pictures should not appear initially on on-line dating sites.  Instead, you have to establish a rapport with someone, and when you’re sufficiently interested, THEN you get the photo.  I’m not going to claim that appearance doesn’t matter–you have to be attracted to your partner.  But it might weed out someone you’re clearly incompatible with first.

I suspect a lot of people don’t want to do this, and they’d be suspicious of profiles with no photos.  The other issue is that when you see that person’s photos for the first time, and you decide you’re NOT attracted to them…. there has to be a way to let the person down easy!  Some kind of automated “Sorry, it’s not you, it’s me” letdown button.   Anyone want to work on that UI?