Address Operational Challenges With Data
March 30, 2020 by
Articles on leveraging data to optimize services or overcome challenges appear daily. Local officials bombarded with such messaging can feel lost in the technicalities. There is desire to leverage an organization’s data and the value is apparent, but how do non-technical leaders go about seeing success in their day-to-day realities?
The good news is that data analytics is not too technical for government leaders to implement. Data is the way policy moves into practice. It helps leaders and their staffs work smarter, not harder. The key is that data must be married to a service change in order to see results. In other words, policy outcomes happen when the insight from data drives actionable change.
To make this a reality in any-sized jurisdiction, leaders must find the opportunities for service change. Think of the truffle supply chain. The truffle pigs sniff and dig out truffles, which have significant and immediate market value. Opportunities to use data analytics are the truffles of local government. Leaders who find those truffles and bring them to market will be heroes.
Following are six practical ways to find data analytics opportunities from Oliver Wise, Tyler’s Socrata Data Academy Director, as shared with attendees at the National League of Cities Congressional City Conference Data Accelerator.
1. Find the Needle in the Haystack
It can be difficult at times to identify or locate service recipients. Data analytics can be used to predict locations or identify targets of a focus so that agencies can focus their resources in precise, non-random ways.
In New Orleans, for example, a tragic house fire claimed the lives of five people. The house did not have a smoke alarm. While the fire department had a program by which residents could request smoke alarms, very few people took advantage of it. The fire chief wanted to prevent future tragedies and proactively send fire fighters into neighborhoods to query households and provide alarms. Using data to inform fire fighters as to where to go first gave them a high hit rate and ensured their time spent was valuable. The city used free, federally available housing data along with its own fire incident data to create an algorithm that predicted block groups most likely to have the highest need. Seven months after fire fighters provided alarms according to the data, 11 people escaped a house fire in a targeted block because of a new alarm.
2. Prioritize Work for Impact
Operational backlogs are common in governments. Often those backlogs are tackled on a first-in, first-out basis. Data analytics can help agencies triage their backlogs so that services are delivered first to those who need it most. This is even more important for chronically under resourced agencies in order to get the most bang for the buck in resource deployment.
In New York City, a small team of building inspectors faced an immense backlog, which included illegally converted buildings that were unsafe for residents. When inspectors visited properties based on complaints, 21% were in poor enough condition to warrant a vacate notice. Data analytics showed that property tax delinquencies were a good predictor of unsafe residences. By simply resorting the backlog to visit tax delinquent properties first, the hit rate increased to 71%.
3. Use Early Warning Tools
Many government services are reactive and used after problems occur. The opportunity here is apparent in that if data analytics can predict problems before they occur, governments can pivot their services to be more preventative than reactive.
Charlotte-Mecklenburg County, North Carolina, instituted an early warning system to mitigate the risk of police officers using adverse force. Instead of being reactive after an unfortunate event, the city reprioritized its resources to deliver services to those officers who were flagged as at risk in order to prevent incidents.
4. Make Better, Quicker Decisions
Some repetitive decisions are made on their own terms, over and over again. Leveraging data to examine prior, similar situations before taking action can provide insight on how to make future decisions better.
With blight a major issue in New Orleans, leaders turned to data to analyze patters in blight abatement decisions. Once a guilty verdict in a blight hearing was reached, the city could choose to demolish the property or foreclose on the lien and sell the property. While the blight operations had been scaled up, the business process around this end decision remained unchanged, leading to a significant backlog. Whether to demolish or sell was a repetitive decision that was made manually, over and over, on each case. Examining the decisions provided insight into patters that led to a blight “coin sorter” algorithm that classified cases into “likely sells” or “likely demolitions.” This reduced an 18-month backlog in just 90 days.
5. Optimize Resource Allocation
Oftentimes in government, assets are deployed based on a hunch or, “it’s always been that way.” Data’s use in this case is to drive decisions on the deployment of resources based on true need, not anecdote.
Chicago, Illinois, reduced complaints about rats by examining the factors that led to rat-related 311 calls. Rat traps had been placed in areas out of habit, but a new look at the leading indicators behind complaints led the city to reallocate rat traps to get ahead of infestation. The algorithm predicted rat complaints would follow water main breaks, illegal dumping or garbage situations, among others, and rat traps were placed accordingly. This reduced rat complaints by 15%.
6. Experiment for What Works
Government sends out mass communications all the time asking residents to take action, from paying a bill to registering for a program. Research shows that very subtle tweaks in how a letter is framed can have significant impact on its effectiveness. Oftentimes, communications are sent across the board without any testing. Employing A/B testing and comparing results can illustrate exactly which messages work so future communications can be adjusted accordingly.
In New Orleans, only 50% of eligible residents took advantage of a new Medicaid program providing primary health care. To encourage participation, the city employed an A/B/C text message test, which showed a clear difference in which wording resulted in significant behavior change and led to more doctor appointments.
Data analytics need not be esoteric; it has practical application and real value in making organizations work smarter, not harder. By identifying opportunities to analyze data, cities have saved lives, increased productivity, averted crises, cut through backlogs, and more.