Using Data Analytics to Assess Factors Affecting the Survival Rate of Young Urban Trees

Urban trees play an important role in cities big and small, both for aesthetics and carbon dioxide sequestration. Therefore, many cities and towns have a community tree planting initiative. The Town of Fairfield, Connecticut, USA has been running a community tree-planting and monitoring project for 6 years. The program plants young trees in the public right-of way adjacent to private homes for a reduced fee. There is a 94% success rate for the trees planted. At the request of the town’s forestry committee, we collected and cleaned the planting data as well as the annual evaluations data for over 700 trees. The cleaned dataset was analyzed for feature importance using the Python coding language. We experimented with three feature importance techniques: CORR, RF, and XGB. We showed that the following factors had a strong effect on the trees’ success rate: tree diameter at planting, tree species, irrigation method, and to a lesser extent, the tree supplier. Also, interestingly, the denser urban parts of the town had relatively higher success rates.