Sergey Nivens - Fotolia

How Boston embraced data analysis methods

The current mayor of Boston is moving to make data analytics a key part of how city services are run, which is a major departure from his predecessor's approach.

Boston's current mayor, Martin J. Walsh, is the first one in the history of the city to have a computer on his desk. He took office at the beginning of 2014.

To be fair, his late predecessor, Thomas Menino, was the city's longest-serving mayor. He began his tenure in 1993, before some people had computers in their homes. Nevertheless, the change speaks to a new perspective Walsh has brought to the office, one that has heavily emphasized technology and, specifically, data analysis methods to solve some of the most common problems cities face.

If Menino was the "urban mechanic," a nickname he earned for taking a hands-on approach to administering the nuts and bolts of running the city, then Walsh might be called the urban programmer. He can come across every bit as interested in engaging with the minutia of city management, like making sure roads get plowed and potholes filled, but he takes a much more data-centric approach than his predecessor that allows his administration to implement fixes at a broader scale.

His technology team built a dashboard for him using Tableau software that allows him to track public safety and crime statistics, information on trash removal and how long building permits take to clear.

"[Walsh is] very engaged with that information," said Jascha Franklin-Hodge, the city's CIO. "It reflects what you see in any large enterprise. There are opportunities to improve and without measuring to targets, you will never see those improvements."

But just like other large enterprises, the city has hit some speed bumps as it's moved toward a modern data infrastructure. When the Walsh administration took over, it inherited a full stack of SAP tools, from data warehouses and data integration tools through business intelligence reporting systems.

The city contains a huge amount of data, and just a portion of that is in a structured data warehouse.
Matthew Mayrldeputy CIO for the city of Boston

Matthew Mayrl, the city's deputy CIO, said these tools work well for what they're designed to do, which is primarily produce static reports from structured data. But he and his team recognize that there is a lot of data in the city's IT systems that does not fit with this structured data world, and using less structured data for new and more complicated data analysis methods takes new tools.

"The city contains a huge amount of data, and just a portion of that is in a structured data warehouse," Mayrl said. "From an infrastructure standpoint, we need a broader set of tools."

The information team has started implementing new tools. In addition to creating dashboards in Tableau, the team has data scientists who do a lot of work in R. The team is also in the early stages of evaluating different cloud storage options.

Franklin-Hodge said R comes in handy when team members want to work on less structured data, like traffic data. In February, the city entered a data-sharing partnership with traffic and navigation app developer Waze. As part of the agreement, the city will share its information on road closures with Waze so the app can alert its users and, in exchange, the city gets access to Waze's stream of traffic information. Analysts are using this information to study traffic flow patterns in order to better time traffic light changes.

Recently, the team used it to figure out a way to reroute traffic around a demolition project to remove an old overpass. A tool like R gives analysts the flexibility to use a stream of data for a one-time project like the demolition without making them go through a lengthy product development process geared more towards putting something into production on an enterprise scale.

"We're not creating new tools from whole cloth," Franklin-Hodge said. "We're trying to take data that already exists and unlock it."

Going forward, the information team will look to pull data from even more sources. Mayrl said the office just implemented a new ERP system that could offer a trove of useful data, and the team is starting to mine call center data and apply data analysis methods.

For example, the team recently ran an analysis of calls to the Mayor's Hotline, a phone service for Boston residents to bring attention to issues in their neighborhood. Analysts looked at unstructured data from calls that did not generate a specific service request. What they found was that the most common reason residents called was to find out when their next trash pickup would be. This led to the development of a mobile app that residents can use to check this information that will come out in the next few months.

"We want to get to a place where we're being more responsive to customer interaction points and using better analytical tools to change how services are delivered based on what people are asking us," Mayrl said.

Ed Burns is site editor of SearchBusinessAnalytics. Email him at [email protected] and follow him on Twitter: @EdBurnsTT.

Next Steps

Help your business users and executives understand data analysis methods

How to develop a data-driven culture

What do you know about data collection and analysis?

Predictive data analysis is more than just a math problem

Read another story about BI tools

Clean data needed for cloud-based analytics


Dig Deeper on Predictive analytics