Have you begun to feel the pressure to start collecting, storing and analyzing data about your residents’ consumer behavior, television viewing habits, doctor visits, left turns? Has the talk about linking resident movements with the tides and sun spots gotten a bit outré? These pressures come from the push for big data where getting lots of bytes and seeing ‘what’ correlates with ‘what’ can help you discover unexpected, interesting and, better-yet, useful information. As the CEO of a San Francisco-based predictive-analytics company put it, “The more hypotheses you can come up with—hypothesis after hypothesis after hypothesis—the more ways you have of finding things that other people don’t find.”
The media examples of private sector data mash-ups with market potential have begun to play loud and obvious like an alarm intended to awaken unimaginative public servants to realize that they are becoming the buggy whip producers of the Model-T age. Maybe you’ve heard that by linking after-purchase car problems with car color, it was found that owners of orange cars treat their vehicles better than do other owners or that by correlating phases of the moon with sales data, one company discovered that deals made on a full moon were 43 percent bigger than deals made when the moon was not full. These real correlations were uncovered by tinkering with large data sets.
Data mashups aren’t just for the private sector. With clever smartphone apps, some cities now harness resident photos of pot holes to target needed street repairs and improve transit service by tracking commuter behavior.
The good news for local government administrators is that data are big, not just because of their volume and the potential for unexpected relationships lurking in the numbers. Data are big because they can answer important questions. Furthermore, local government does not want for large volumes of useful data even in the absence of any smartphone apps. Local government does not need movement sensors or satellite downloads to become a big player in big data. Government needs thoughtful leaders and skilled researchers to determine what to do with the data it already has.
In a workshop my colleague recently gave to city managers in Nebraska, she identified scores of medium to large datasets that most local governments already store. For example, police already have decades of crime data. Code enforcement and fire services already have years of violations data. The utilities department has water usage and video monitoring of sewer lines. There exists gigabytes of records about historical media circulation from the library, recreation facility use from the parks department and more.
So forget worrying about finding the next set of large data. Data are like grains of sand: in large quantities they just pile up. You wouldn’t order tons of sand unless you knew if it was for a beach, roadway, construction, artwork, glass-making, ballast or an hourglass. Similarly with data applications, someone must identify a use, ask what else is going to go with the sand and then the mashup can create the cement.
In part because of “silo-based’ systems storage, failure to designate anyone to generate questions and no assignment of anyone to crunch the numbers, many important questions that can be answered with local governments’ big data are never even asked. Data that reside in local government repositories are just mounds of sand waiting to be sculpted into castles. Local government leaders don’t need more data. They need to assume the responsibility of asking the right questions so that the data they have can be connected to give answers the community needs.
To help them, local government chief administrators should empower department managers to work together to ask the penetrating questions. Managers should hire savvy researchers or evaluators (who are different from information technology engineers), those skilled in helping to hone the right questions and with the computer skills to get the answers.
Check out this document if you want some help picturing the data you already have and the questions about those data that may turn up unexpected and useful answers.