Nothing says January like a list of resolutions calling for healthier living in the new year. I have admittedly given up making such lists for myself, but I’m happy to say that I played my part in promoting good health by getting my wife an activity tracker for Christmas. She will join 40 million other Fitbit, Jawbone and Apple band wearers in tracking multiple aspects of their daily activity and sending loads of data back to device manufacturers in 2015.
The Activity Data Gold Mine
At the level of the individual wearer, the utility of an activity tracker is quite evident. It tracks raw stats like heart rate, steps taken, distance traveled, calories burned and location, among others, allowing users to monitor, visualize and brag about their daily activity.
What may be more intriguing is the data these device makers are sitting on — a veritable gold mine, one that grows deeper with each passing moment. There is significant value to be extracted from the aggregation of data across millions of activity tracker users. A simple and timely use of aggregated data might be a visualization of New Year’s Eve activity split out by age or by country of residence, but more valuable applications abound.
More so than analytics in most other fields, the insights from activity data can be literally life altering. It is through the analyses of data from substantial numbers of activity tracker wearers that researchers can hone the accuracy of personalized recommendations: suggestions to increase activity; changes in sleeping habits; or ways to calm down. A rather more ambitious goal, and one with great potential to impact quality of life, is the early detection of disease through the use of activity tracker data. These are early days yet, but I am optimistic there will be many novel analyses of activity tracker data in the near future.
A Textbook Big Data Problem
The extraction of useful insights from activity data is a prime example of a big data problem. With millions of activity trackers in use, each housing multiple sensors throwing off telemetry data 24/7, device makers are collecting volumes of time series data that need to be stored for analysis. In addition, big data tools are more adept than traditional methods at handling the real-time and concurrent nature of the data collection.
One virtue of this being a relatively young product category is that there are no legacy data warehouses to speak of, which makes the adoption of big data technologies much more straightforward. Indeed, a browse through several device-maker career pages reveals keywords like Hadoop®, Hive™, Pig, HBase, Impala, Spark™ and Shark, all software widely used for large-scale data processing. And even if you aren’t familiar with all these tools yet, just resolve to get healthy for 2015, slap on that activity tracker, and you’ll be contributing to big data in no time!