Could the Future be “Small Data”?

July 17, 2019

Self-tracking of personal data is on the rise with 69% of Americans keeping track of some kind of health data. Twenty-one percent of the population use some type of wearable daily. Forty percent of adults in the U.S. say they foresee buying some kind of wearable device. Quantified Self, a community for people who self-track, has meetups in 119 cities in 38 countries and regular international symposiums.

Most people report tracking physical activity, food, and weight. Thirteen million wearable activity trackers have been deployed or subsidized as part of a corporate wellness program. Ten thousand companies worldwide have offered their employees fitness trackers. 110 million wearable sensors were shipped as of the end of 2016 (Neff and Nafus 2016).

Venture capital funding in the self-tracking space doubled since 2017. Furthermore, crowdfunding sites are rewarding quantified self tools. The Misfit Shine, an activity tracker, raised $446k. uBiome, a tool to sequence microbiomes, raised $351k. Pebble, a watch that tracks activity, raised $10 million. Amiigo, a fitness tracker, raised $580k. These products are showing up at the Consumer Electronics Show as well as SXSW (https://technori.com/2018/08/4281-the-beginners-guide-to-quantified-self-plus-a-list-of-the-best-personal-data-tools-out-there/markmoschel/ ).

With all this data accessible to the layperson, is the future of data analytics “Small Data?” Now, with so many devices tracking personal data, I think that there will be a need for non-specialists to crunch their own data.  Just as mainframe computers used to take up a whole room and require PhDs to program them, “Big Data” requires teams of experts to evaluate reams of data.  I think the future will be laypeople crunching their own small data sets much like people programming their own personal computers.  The key is to make data analytics fast and accurate.

I mocked up a website that would allow anyone to run a regression–a data analytic tool that predicts the relationship between two variables. For example, if I had data on weight (in pounds) and calories consumed, I could plug in my data to the website and find out how an increase in one calorie consumed leads to how many pounds increase.

One person I interviewed, Michael, had this to say about how he tracks data and whether he would use a tool like regression analysis:

“Michael:  Yeah.  I can totally put data in there.

Helen:  Well, what do you use to manage your data?  Excel?  Or?

Michael:  Well, it’s a hybrid.  I use MyFitnessPal to keep track of my food.  My FitnessPal, it’s not the most intuitive app, but it’s all I have at the moment.  They have a robust data base of food already loaded in so I don’t have to enter new food.  I also have excel, so with that every day I weight myself when I wake up in the morning after I use the restroom.  I track that every single morning.  And every day there’s a tracker I fill out that tracks if I stayed within my calorie range.  Did I meet my step goal for the day and then how well did I sleep, how hungry was I, how stressed was I, and how much energy per and a section for notes.  And in addition to tracking my weight every day, once a week, I take measurements in the morning along with my weight.  So I measure my arms, my chest, my stomach in 3 different areas and then my thighs.  And that’s pretty much it.  That’s what I track everything with.”

From my interviews, I discovered in what scenarios might someone use a data analytic tool. Here is a list of items my interviewees mentioned:

  1. Health—weight loss and calories, blood pressure and salt intake, carbs and weight loss
  2. Sales people and productivity
  3. Gas levels and miles driven
  4. Number of pharmacy scripts fulfilled and revenue
  5. Number of active users and revenue
  6. Hours spent preparing for lecture and student satisfaction
  7. Hours spent writing journal articles and likelihood of publication
  8. How much you feed a baby and how many hours she sleeps
  9. Number of conference room square footage and how much productivity
  10. Return on ad spend

For the data they already keep, my interviewees mentioned that they use Google Drive, Google Sheets, and Google keep because it’s free, it’s accessible on any device, and it’s shareable. At work, they use Excel. Currently, they use Google Analytics to crunch their data at home and Tableau and proprietary software at work. At work, they can also hand the data analytics over to the business insights team.

Producing an analytic tool that is easy to use is not without hurdles. I found I had to do a lot of hand-holding to get my interviewees to use the website correctly. They had basic questions such as, “What does the analytic tool do?” Also, “How do I enter data from my spreadsheet to the analytic tool?” Also, “Is it compatible with Google Sheets, Google Drive, and Google Keep?” Finally, they needed guidance on how to interpret the output.

If we can solve the problem of automating regression analysis and make it easy, fast, and accurate, I think the future of data analytics may just be “Small Data.”

Leave a Reply

Your email address will not be published. Required fields are marked *