DIRECTOR'S ANALYSIS: Hiring challenges
By Stephen Goldsmith and Katherine Hillenbrand
Data analytics is such a rapidly-advancing field that the demand for top-tier analysts far outpaces the supply, driving up salaries. Given the limited funding available for city employees, municipal governments should turn to the competitive model that commercial enterprises have recently adopted to attain data insights that would otherwise be out of their reach.
Companies from Amazon to Walmart are capitalizing on the valuable knowledge that complex data analysis can provide about consumers, including the now infamous ability of Target to know when a woman is pregnant before she even tells her family (New York Times article). Such intricate analytic capabilities offer even more public value when applied in local governments to everyday concerns like transit operations and public safety. Of course, the deep pockets of large corporations dwarf the shrinking budgets of struggling cities. Despite the promises of rapidly-advancing data technology, most municipalities can’t even consider an expensive new proposition like hiring a permanent, highly-skilled data analyst.
A new private-sector model could provide a feasible entry point for cities looking to become more effective and efficient through analytics. Kaggle (tagline: “making data science a sport”) hosts funded competitions from multifarious companies, including BestBuy, Allstate, and Facebook. Each posts a challenge and accompanying data sets; statisticians and data scientists from all over the world compete to develop the most accurate model in order to win a set prize. Prizes range from $3,000,000 for developing a model that predicts which patients will require hospital care in the next year, to a Facebook job interview for predicting which other users someone would be likely to want to follow. The competitive nature of these challenges incentivizes continuous refinement of models over the course of each as various teams vie for the lead.
The success of programs like Code for America proves that the concept of public service still wields clout. Data analysts would likely enter competitions to help local governments for a much smaller prize, given the additional altruistic reward of serving the public good. Furthermore, grant funders might be willing to provide the reward to spur participation. Different cities could unite around a common concern and pool resources and data to host a Kaggle challenge, which could be designed to produce models relevant to multiple municipalities. The combined brainpower of the most advanced data analysts around the world may very well produce the analytic models that will change the way local government works.
-----------------------------------------------------------------------------------------------------
Stephen Goldsmith is the Director of the Mayoral Performance Analytics Initiative and the Daniel Paul Professor of Practice at the Harvard Kennedy School of Government.
Katherine Hillenbrand is a research assistant for the Mayoral Performance Analytics Initiative.