Kenshins using data science to visualize the future of Boston’s urban tree canopy

Caroline Gerenyi
Kensho Blog
Published in
4 min readApr 16, 2020

--

As part of Kensho’s 2020 community impact initiative, Kenshins have been volunteering their data science skills to help a Boston-area organization called Speak for the Trees, Boston (SFTTBos) to improve their collection and analysis of data about greater Boston’s urban forest.

Speak for the Trees is working on a multi-year volunteer and resident-led project to conduct a comprehensive census of all street trees in Boston. This involves documenting the location, planting site characteristics, size, species and markers for disease for all of Boston’s street trees. Similar tree inventory work has been done in New York City, with a tree count once every ten years, which Boston has been using as inspiration for this project.

The above visualization shows the tree diameters in inches for the Roxbury neighborhood, sorted by species.
The above visualization shows the tree diameters in inches for the Roxbury neighborhood of Boston, sorted by species, one of the many tree metrics that SFTTBos tracks.

Kenshins Hamima Halim and Sireesh Gururaja learned about the organization and their need for data science help at an event at Impact HUB Boston, a community and co-working space for businesses addressing important challenges.

Hamima and Sireesh met with Speak for the Trees’ Co-Founders David Meshoulam, Executive Director, and Amanda Rich, Program Director, to better understand their needs and what they’d like to see done with the data. For example, which factors in the data might show that a tree could be in trouble in the future? Which kinds of trees are no longer being planted in Boston? What might the canopy look like without these trees? How large and healthy do various tree species grow in the city?

One specific area in need of analysis was the reliability of the field data. The census data was collected mainly by a group of teenage volunteers, and a subset of it was field-checked by a trained program coordinator every two weeks. Over the course of a summer, the teens were able to canvas multiple neighborhoods in the city and counted almost 4,000 trees, as well as identifying over 600 sites where the city could plant new trees.

The data was complicated to make sense of in its raw form, so they began by putting together some visualizations to pull out some key insights. A few of these charts are shown below.

Side-by-side comparison of all and compromised trees in the Roxbury neighborhood
Above is a side-by-side comparison of all trees in the Roxbury neighborhood. On the left, we see red dots marking species of trees that are currently threatened by factors such as the emerald ash borer, and on the right, what the canopy would look like without the at-risk trees.
Ever wondered what the most common tree types were in your neighborhood? This chart shows the number of trees by type for one neighborhood in Boston.

Accurate tree data is critical to all aspects of the project, including fundraising and engagement with municipal officials, as donors and arborists want to be assured that they’re seeing a clear picture of both the current status as well as signs of improvement to the urban canopy as trees are planted and cared for. How could the team convince fundraisers and stakeholders to trust the volunteers? With a fundraising event coming up on March 24, answering that question was important and time sensitive.

Hamima and Sireesh got to work on their data crunching. QA data had been collected in the form of biweekly spot checks with notes from the coordinator on changes made. The notes needed to be parsed, and failure modes needed to be categorized and tallied. Upon analysis, the data collected by the volunteers and teens was very accurate. Although there were occasional errors in the reporting, on the whole, the teens’ reporting accuracy ranged between 87–98% across all data fields. The project is in great hands with the teen scientists!

the accuracy of the work being done by the volunteers for SFTTBos
The accuracy of the data being collected about the trees is critical to the success of SFTTBos. The above visualization shows the accuracy of the work being done by the volunteers.

Speak for the Trees will continue to collect data for their projects moving forward, helping answer the important question, “What will the future canopy of Boston look like?”

In addition to running tree inventories, Speak for the Trees does advocacy work encouraging communities to grow and care for more trees and holds tree giveaways several times a year.

Their inventory work will continue until the leaves fall. You can learn more at sfttbos.org, donate at sfttbos.org/donate, and follow them on FB, Twitter and Instagram @SFTTBos.

--

--

Caroline works for Kensho Technologies in their Washington DC office.