How Happy Are the #100HappyDays?
Ah, happiness. Who doesn’t love to be happy? But what is it that lies at the true root of our happiness? Is it family? Is it accomplishment? Is it amazing, life-changing music? Well, a new trend that is sweeping social media might be able to give us some answers. It’s a trend that challenges individuals to be happy, and talk about their happiness—every day—with their social circles for one hundred days in a row.
The #100HappyDays trend started on January 1st of 2014 and continues to go strong. Joining the effort is painless—participants simply choose a starting date and begin posting at least one social post per day with the hashtag #100HappyDays. In that same post, people also include a hashtag indicator for the number of days they’ve been thinking about their happiness (starting with #Day1.)
With almost two million mentions of #100HappyDays on Twitter alone, we have a pretty good data set to dive in and figure out what is making people happy. But beyond that, social data can also tell us a few more things about social experiments like this one. In particular, I’m interested in learning:
1) What’s making people happy?
2) How long are people sticking with the #100HappyDays
3) How happy are people during the 100 days?
Luckily, we have some social data lying around here at Spredfast. Let’s take a look at the patterns that the #100HappyDays hashtags generate and see if the resulting insights put a smile on our faces.
I looked at Twitter data since January 1st of this year for overall patterns of people using the #100HappyDays hashtag as well as hashtags representing progress over the one hundred days. I pulled volume and sentiment at each ten day increment, meaning we ended up with data for #Day1, #Day10, #Day20, #Day30 and so on.
A High Level View of Happiness
Sometimes when we take a step back and look at data from a 30,000 foot view we can see some cool patterns emerge. Let’s take a look at what we have to work with:
Woah, that’s a big, happy mess. There’s a lot going on in this graph—too much, in fact. The only takeaways that I get from looking at all this data are 1) we need to look at less data and 2) there’s something interesting going on in the bottom left.
Let’s break out a few of the daily progress hashtags into individual graphs to see if we can find a pattern here.
Well, now we’re on to something. As we march through each hashtag, we see something interesting happening. The pattern looks very similar for each day’s hashtag, but just pushed back roughly—wait for it—ten days each time. This makes total sense. As people continue through the #100HappyDays challenge, they follow the same posting patterns as people that started on the same day as they did. That’s why our view of all the hashtags is such a mess—it’s like trying to play every minute of a song at the same time, it’s just going to sound like noise. So how do we get rid of the noise?
Cohorts—Happy Little Buckets
We’ll use a technique called cohort analysis to make sense of our happy data. A cohort analysis helps us arrange individuals into groups that share a common characteristic and uncover insights once our data is lined up in a new way.
Remember—one of the things we’re trying to figure out is how many of the people that started the #100HappyDays challenge actually finished the journey. To figure that out, we’ll want to use cohort techniques to group people by the day they started, and analyze the retention (and dropoff) those people experienced as they walked through their one hundred days.
To start getting happy, let’s slice our data to look at everyone who started the challenge on a random day (Feb. 12th), and watch that group’s progress through the challenge. The resulting retention looks like this:
So our cohort analysis tell us that roughly a third of the people that started on Feb. 12th finished with a big #day100 post. Strangely—we also see that between #Day30 and #Day40, the number of people posting progress went up. Cheaters.
The fun thing about cohort analysis is that we can now compare that dropoff rate to people that started on another date. I’ll pick another random start date (March 9th) and see how the two groups compare with how they stuck with the challenge:
So our group that started on February 12th did much better than the later group—twice as well, in fact. 30% of the people that started the challenge on February 12th finished, while only 15% of those on March 9th completed the journey.
Now let’s use this technique to look at retention patterns across all of our #100HappyDays data. Making a group for every day of the year would get confusing, so I’ll bucket participants into the month that they started, and watch their dropoff rates since then. Will people that started the challenge at the beginning of the year stick to the program with more oomph?
As you can see, the overall pattern stays pretty consistent. Participants that began the program in January did the best in sticking through until the end, with an 34% average completion rate. Every other month has seen pretty much the same shape—ending up with around 25% of the original participants staying happy (and social) all the way to #Day100.
What Makes You Happy?
So there’s more data in here other than just completion patterns—we can try to gauge how happy people actually were as they participated in the #100HappyDays exercise, and what exactly they are happy about.
I used Spredfast’s shiny new Intelligence tool to mine the sentiment of #100HappyDays Twitter posts by day in an attempt to detect patterns of happiness throughout the experience.
Which of the 100 happy days was...the happiest?
Day 40, closely followed by Day 1, marks as the day where participant’s #100HappyDays posts have the highest positive sentiment. The good news is that the average positive sentiment from participants on every day was higher than the normal level (5-6%) seen across Twitter. In other words, the happiness expressed from users’ #100HappyDays Tweets was indeed happier than the average Tweet. Phew.
And not only can we see how happy people are, we can also see what topics and subjects they are happy about. Because part of the #100HappyDays exercise is to talk about what is making you happy each and every day, we can simply mine that data to discover which subjects were mentioned in the highest frequency.
There’s nothing quite like love (and apparently, no place like home) to make people smile. Friends and coffee aren’t so bad, either. Take that, selfies.
A Few Final Happy Thoughts
In the social media age, trends come and go faster than ever. Asking anyone to commit to 100 days of anything on social media is a tall order, and while the majority of participants didn’t finish the #100HappyDays, it’s great to see that many are sticking around to finish what they started.
So tonight when you head home, don’t forget to tell someone you love them, call an old friend to say hi, and maybe even sneak a little piece of chocolate. You can say the data made you do it.