Investigating Sleep Patterns During the 2017 World Series

Benjamin Bradshaw
Feb 16, 2018 · 5 min read

In the spirit of baseball season starting, the data science team at , a division of , decided to look back at the World Series and do a deep dive into sleep patterns.

The World Series is an interesting athletic event because it spans the better portion of a week and games are typically played in the evenings. When games go long, often times baseball fans stay up late into the night watching the game, especially when the games are exciting. That’s the theory at least, and the Evidation Data Science team was interested in digging into the data to see if we could see changes in sleep patterns during the World Series.

One of the really cool things about the kinds of data that the we work with is that it affords us the ability to infer changes in population level behavior acquired via signals from individuals wearing activity trackers. For this analysis, we used data from a set of Achievement members who were actively using a Fitbit during the one month period surrounding the world series.

Comparison of Time Spent in Bed

One of the things we were curious about investigating was whether there were differences in sleep patterns between Achievement members who live in World Series cities (e.g. Houston or Los Angeles) to see if we could detect changes in the amount of time spent in bed. The theory being, you are more likely to stay up and watch the game if your home team is in the World Series.

Luckily, in order to test this hypothesis, we are able to observe minute level activity data supplied via user’s Fitbit devices. This data allows us to infer a number of things about a user’s sleeping behavior - one of which is how long they spent in bed on a particular night.

The figure below shows the normalized amount of time users spent in bed during the period of time surrounding the world series. The orange line is the average time spent in bed for Houston/LA users, while the blue line is the average time spent in bed for Achievement users in other cities. The first vertical dashed line shows the first game of the World Series and the second shows the last game.

You can see that the orange line begins near the blue line, but around the time the world series begins, the lines begin to diverge. Achievers are sleeping less than they normally do in LA and Houston when the World Series is hpapening compared to Achievers in other regions of the country! A pretty neat result!

Do sleep habits vary by gender across locations?

Up next on our list of items to investigate was the question of whether sleeping habits during the world series varied by gender. In order to do this we split the same group of users into male and female cohorts, and then subdivided each gender subgroup again by location (e.g. World Series location vs not).

Each of the lines in the figure below show average normalized in bed minutes by date during the period surrounding the world series for a specific subgroup. The orange and blue lines represent non world series region gender segments. You can see that the users in those regions see very little deviation from normal during the World Series.

The red series shows men who live in either LA or Houston. Interestingly, men seem to be sleeping less than normal before the world series even begins. Their time in bed rebounds to normal during the weekend but then plummets on the night when the last game of the series took place.

Comparing that to the green series which represents women in World Series cities, we see that for the most part women are not losing sleep during the world series until the last few days when it got really exciting. During the last day, women in these cities lost even less sleep than the male-LA/Houston cohort. Interesting!

How does having children impact sleep behavior during the world series?

The last item we wanted to investigate was whether having children had an impact on time spent in bed during the World Series. In order to answer this question we followed the same procedure as the gender investigation: split the groups into people with and without children, and again divided these groups by the relevant regions.

Again, we see that for non-World Series regions, there isn’t much of an impact. Things get more interesting when looking at the red (LA/Houston users with no children) and green (LA/Houston users with children) series. We see that the red users consistently slept less than normal during the world series, indicating those without children in game regions were staying up later to watch the game.

That is not true of the green (game region-have children) users. These users slept normal amounts during the majority of the World Series, and then the last night of the series saw a large drop in the amount of time they spent in bed. Perhaps the excitement of a historic game allowed the kids to stay up late with their parents?

So to answer our original question: did the 2017 World Series impact sleep patterns within the Achievement population? The answer seems to be yes, although the effect appears to vary quite a bit depending on geography, gender, and if the users have children.

Are there sporting events or awards shows that make you stay up just a bit later than normal? Let us know below.

Achievement - Support your health. Reward your life.

Benjamin Bradshaw

Written by

Data Scientist @evidationhealth

Achievement - Support your health. Reward your life.