Skip to content

@BARIexplorer beats the heat: tweeting the temperatures of Boston’s streets

Summer is here! And with it, hot weather. Heat is more than an excuse to go to the beach or sweat-inducing inconvenience, however. It can be a major health hazard, especially for the very old, very young, and those with underlying conditions. To this effect, starting today we are repurposing the BARI Parcel Explorer—our Twitter bot that tweets out information about a random parcel in Boston every half-hour (@BARIexplorer)—to tweet about temperature on street segments across the city.

Why street segments, you ask? BARI recently published a paper in American Journal of Public Health examining temperature and medical emergencies across Boston during heat advisories (defined as days with a heat index of at least 95°F). We took an environmental justice lens, asking whether certain communities are exposed to more deleterious conditions than others.

There has been plenty of research on the “urban heat island effect,” but we framed it in a new way. Drawing on insights from criminology that crime tends to concentrate at hotspot streets, we wanted to know if the medical consequences of elevated heat are not just a matter of urban vs. rural, or even about specific neighborhoods, but could in fact be about hyperlocalized contexts, defined in this case by street segments. The results were two-fold:

  1. There was substantial variation between neighborhoods in temperature, but there was also a considerable range in temperatures between the streets of the same neighborhood. For example, the figure below shows the whole city and an inset of a single census tract in northern Dorchester that has very hot areas and quite cool ones.
  2. Variations between streets and not neighborhoods predicted the likelihood of medical emergencies during a heat advisory. (This finding took account of the tendency of a street to generate medical emergencies during cooler months.)

The takeaway, then, is that the city is not a single “urban heat island,” but more of an archipelago with many smaller “urban heat islets.” Further, these urban heat islets are the key to understanding how high heat impacts public health.

The bot’s newfound purpose is to tweet out one street segment every half-hour, describing its land surface temperature, land use, and neighborhood, accompanied by a map. We hope this serves as part as a public service announcement, but more to draw attention to nuanced geography of heat emergencies and public health. If you would like to explore the data further, please visit the Boston Data Portal to view or download the Urban Land Cover and Urban Heat Island Effect Database yourself (made by recently-minted PhD Dr. Andrew Trlica, with support from a BARI research seed grant).

There are a few important methodological notes to mention as well. The temperature data, as noted, are from USGS’ LandSat data, which were collected for 30m grid cells over the course of many summer days. We then spatial joined the grid cells to street segments and used a weighted average (based on the % of the street segment falling in each grid cell) to estimate its temperature. Importantly, these data were collected 2002-2008, which means they may not be a perfect representation of current temperature on the ground. Many neighborhoods have largely the same structure and, thus, a similar temperature now as then, but others (e.g., the Seaport) probably are a bit different. Also, land surface temperature is quite a bit higher than the air temperature we experience, meaning that some of the numbers seem a bit ridiculous (upwards of 110°F). Last, we measured medical emergencies from 911 dispatches, which are more sensitive to patterns of reporting than, say, hospital records.

Those caveats aside, the findings and the data hold much value for the city, especially during these hot months. No matter how you want to engage—by following the bot, visiting the map, or downloading the data to do more analyses—please enjoy, and stay cool!

More Stories

How Does Vaccination Unfold across Communities?


Mental Health Effects Varied with Sociodemographic Characteristics


From Inequities in Vaccination Uptake to Disparities in Herd Immunity