Courtesy of Adam Storeygard
Economics research is drawing upon new types of data, new research methodologies, and considering new populations. Urban development economist and Tufts Economics professor Adam Storeygard is exploring global trends of vehicular speed in cities, particularly in India, with a metric he calls the speed index. He and his team use Google Maps data to determine travel times between locations in a city and compare these travel times to equivalent trips in other cities using regressions.
“Just like you can ask Google Maps how long it would take you to get to the airport from wherever you are right now, we basically do the same thing, except we do it about 50 million times in 180 cities,” Storeygard explains.
By asking Google for the speed of each trip with and without traffic Storeygard and his team could distinguish the effect of congestion from non-congestion factors, such as the slope and width of a road or how well it is paved. They concluded that non-congestion factors are more influential on speed index values than congestion factors.
“We found [this conclusion] surprising because people think of lots of big cities in India having lots of congestion... it’s true that there’s a wide range of congestion, but there’s also an even wider range of uncongested travel,” Storeygard said.
Each Google estimate, and related ones from other sources like Uber, are generated from the travel patterns of many phone users. Storeygard’s team found additional data on the actual travel of individual vehicles thanks to an app developer in India who was collecting had data on a specific non-congestion factor: potholes. When a car rolls over a pothole, a significant jolt is recorded by the phone’s accelerometer. These data allowed the team to verify Google’s estimates with the travel of real people.
Storeygard’s interest in geographic information systems (GIS) data, which includes examples like these, dates back to internships using such data during college, when it was less widespread than it is today. After earning an undergraduate degree in physics, Storeygard pursued a master’s degree in geography with hopes to travel the world, leading him to study in the UK, where he honed these skills further. As his master’s degree exposed him more and more to the problems influencing the developing world, Storeygard gained interest in topics of development, introducing him to the prospect of specializing in economics. Ultimately, he was torn between geography and economics, eventually choosing the latter due to his interest in the causal thinking mindset that is essential in econometrics and quite applicable to social science research questions.
While the team is in the process of revising their paper, Storeygard has plans for future research. Although speed itself is not an indicator of welfare, speed has important implications for accessibility to both local and distant amenities like stores and jobs. Thus, Storeygard and his team hope to do analyses on accessibility in the future, a more challenging project, yet one of great importance in the field of development economics.
“It’s not like everyone is better off if everyone drives faster. But, we do think that accessibility is an important goal... accessibility is the thing we care about for welfare, but speed is an important component of accessibility,” he explained.
The benefits of Storeygard’s research experience have spilled over into his life as a professor and mentor. He can update course material regularly, such as in EC-35: Economic Development. As a mentor, he pulls upon his research experience to give pointed, up-to-date advice on Ph.D. students’ research prospects and set them in the right direction. He notes that research is not for everyone, so gaining some research experience before committing to obtaining a Ph.D. will ensure that the five- to six-year journey as a Ph.D. student will remain enjoyable.
In reflecting upon his work, Storeygard is enthused about being able to find answers to novel questions, not only by calculating speed index values to cities in India but by implementing an unprecedented scale of data aggregation and analysis from over a thousand cities worldwide. The accessibility of new data sources, such as Google Maps satellite data and sensor data in cellphones, has democratized economics research by contributing data that is comparable across both developing and developed countries. While the vastness of this data poses challenges to analysis, it has also provided Storeygard with opportunities to experience the increasing level of collaboration that is characterizing the field of economics.