Enforcing Information Policies Using a Small-Data Approach

People love Uber, the car service app. Setting aside the ongoing discussion over its legitimacy, people have fallen in love with the technology. Uber’s app displays a map, showing the user’s location and the GPS location of the car that they have just ordered, enabling them to monitor its progress toward the pickup location and estimated arrival time. The innovative user-centric design of the app and ability to monitor GPS data, displaying it in an easy-to-understand fashion and thereby setting expectations, have won people over as loyal customers. In our personal lives, we are surrounded by many more apps that can analyze and display other data sets. Mint can access your bank accounts and provide insights into your spending habits, and also lets you set budgets and goals. More interestingly, Mint automatically sends out alerts for anomalies, such as high spending for certain categories. Nest and Fitbit take it even a step further, by monitoring and collecting data from your physical world. Nest is the thermostat that learns from your behavior and automatically sets schedules and can detect whether you are home and adjusts itself accordingly. At the end of each month, the app sends out a report showing energy usage over time and explanations why your energy bill might have been higher or lower, thus providing actionable information to make a decision. Finally, Fitbit is the latest rage for helping people to get in shape, by monitoring activity and exercise through a small wristband. Certain models of the device track steps, miles, heart rate and calories burned. Weekly progress reports can motivate the user to hit certain goals and stay on track on a fitness program. Small alerts provide positive reinforcement to the user when he or she is close to finishing a set goal.

It is clear that we have accepted data-driven tools and technologies to help us manage our lives. These apps have one main thing in common: They seek to solve a particular problem using a limited amount of data. Rather than big data projects, which seek to analyze vast quantities of data to recognize patterns within that data, the data we use to make a decision on whether we need to take more steps toward our daily goal or adjust the thermostat are much smaller and could be coined “small data” tools.