What Statistics and Analysis Tell us About Love
This Valentines Day, I want to tell you about my love for people…watching. You can call me a creep but you’d probably have to call yourself a creep then too. I love the little differences between how people should behave and the way people actually do. I live everyday to enjoy these little morsels of idiosyncratic human behavior. You know what I’m talking about. It’s the moment you catch a guy at the gym just staring at a girl’s butt as they walk by or when you go to a concert and see your boss dressed as a centaur.
Internet analytics give you better insight into human behavior than any kind of observation or report of everyday life. If you want to know yourself, skip a psychologist and keep your browsing history and take a look at it over the last month. You are your browsing history. View your history since 1990 and you will get a better “timeline” of your life than Facebook could ever hope to portray.
We cruise the Internet, like we cruise around our own homes, not wearing pants and eating scrambled eggs over the sink. Ok, that was my morning. We all tend to wear a scruffy old pair of sweatpants more often than we probably would on a first date. You could call it self-awareness, you could call it shame, but the fact is, we act differently when we are not around anyone else. Our use of the Internet shows our own intents and purposes in this very same way.
It’s Valentines Day today and people are looking for love. Well at least they’re searching for it. What better day is there to share with you what analytics has told us of the enigmatic human behavior of love. Ok, and lust too.
They’re searching for it. See, Google says so:
Analytics is all about quantifying data and finding meaningful relationships within the data. It is much like your everyday use of intuition but we base our decisions on large collections of data. Internet usage, and it may surprise you, is made into metrics and quantified. Quantified metrics, now things are really getting sexy.
One common metric we use is time. Just like you use time to measure things in your relationship, we use it to:
Aggregate interest of webpages and activities. Like when you:
Check the duration of a phone call with your significant other.
Measure rate of change of something. Like when you:
Note how many times you have to ask your significant other to do something for you, per time, over time.
Measure trends between variables. Like when you:
Notice how much your significant other will do your dishes for you proportional to the amount of time you do “X.” No pun intended.
Here is an example of how we use time to measure a relationship of time. To clarify, we segment each age (a measurement of time) by each whole year and by the ages of women. Then we measure what ages these men are pursuing. That’s why there are distinct pixels in this graph. Each pixel box represents the proportionality of occurrences the whole population of a man of that age Y is sending a message to a women who is age X. As you would expect, it is very uncommon for a 22 year old man to send a message to a 45 year old woman, not proportionally uncommon for 45 year old men to message 22 year old women.
To be fair, I will show the data for both relationships of the sexes.
Another metric we often use is location. We use location to target specific audiences and to see what audiences are searching for and consuming our content. I was happy to fumble across this article from the Huffington Post during my research. As it turns out Portland, OR is having the most promiscuous sex over any other location. Portland seems to have a lot of notable distinctions. For some reason, they are all the wierd ones.
The breakdown looks like this:
5. San Francisco
7. San Bernardino, Calif.
9. San Diego
We measure this data in the same way that we made the heat map of message target rate (above). We use location coordinates rather than age coordinates to show the expression of a variable. In this case the variable is frequency of bathing.
The question is: How often do you bathe or shower?
I don’t know what this says about love, but again it sure gives us Oregonian another dubious distinction.
Another way we measure human behavior is by quantifying content. If you have ever had a girlfriend, you know that they count “I love you,” and their margin of error is probably less than one. Google does this too in several ways. It measures quantity of content searched for and consumed for trends (above example of “love” and “valentines”) and it measures quantity of content created to show relationships.
Here is an example of how we can quantify specific terms and visually show how they relate to a second variable, in this case, message reply rate. Creepy or not, Ok Cupid has “read” all the text of a lot of messages and found that these terms have a special value.
Here we see how mentioning specific interest and characteristics relate to the reply rate to a first message. Consider that the overall average reply rate is 27%. This would suggest that people might:
a) Have significantly higher interest in these things
b) Place a high value in commonality of these interests or characteristics
c) Recognize that the portion of people that value these things is relatively low so the value of having them in common is relatively high
Finally, because I love flow charts.
We can use page analytics to show sequences. Most often we measure navigation, but in this case, this is a sequence of conditions that would determine whether you are a match for OK Cupid’s creator Chris Coyne.
In conclusion, I love that everyday Zoom Creates pays me to geek out about human behavior. It’s not always as exciting as love and lust, but it is always about relationships. I hope that you have enjoyed this and maybe that math, statistics and analytics became a little more interesting and relevant. If you are one of my friends… and you made it all the way through, hopefully now you can understand why I am so weird…besides living in Portland.
Follow us for more great stuff about design, website development, and analytics.