Embedded Files
AI for Urban Greening: A Critical Analysis
AI for Urban Greening: A Critical Analysis
In Baltimore, a vacancy crisis and continuing impacts of historical ‘redlining’ that discouraged investment in minority neighborhoods prompted city officials to use an algorithm generated Housing Market Typology (HMT) map to guide municipal service distribution. The current strategy channels maintenance to wealthier areas and prioritizes tree-planting and demolitions in struggling areas.
This HMT map groups census blocks based on housing market status and demographic information and is used to make recommendations for municipal service prioritization. Housing activists in Baltimore refer to the map as “Redlining 2.0” because this method is reminiscent of HOLC practices. The different grades are prioritized for different services and investments. We’ll focus most on the grades I and J (shown in orange) which contain the most vacant properties and lowest sale values and many red and yellow lined areas. These census blocks are targeted for demolitions and greening initiatives like tree planting.
Due to the significant overlap between HMT’s ‘struggling’ areas (filled with bright red and orange) and historical ‘redlined’ neighborhoods (outlined in red), citizens are concerned that HMT will increase racial bias in city management.
The strategy was effective at increasing tree numbers within redlined communities, but tree health is disproportionately low in those areas.
Methods
Methods
I used data from the Baltimore Tree Inventory to calculate tree health and density within HMT and ‘redlining’ boundaries.
The Home Owners Loan Corporation ('redlining') data came from University of Richmond’s Mapping Inequality project which is a comprehensive map on redlining in the US.
The 2017 HMT dataset was purchased from PolicyMap, a mapping and data analytics platform.
I trimmed off the HOLC areas that lie outside current city limits, and then I found the percentage of each type of tree (stump, dead, poor, etc.) inside of each type of HOLC area and each type of HMT block.
Results & Conclusion
Results & Conclusion
Good quality trees are less prevalent in Baltimore’s struggling housing markets. There is a 12-15% percent decrease in the average proportion of good quality trees when comparing the upper and lower ends of the Housing Market Typology spectrum.
This reflects the municipal investment strategies that target tree planting efforts to these communities to attract private investment but channel upkeep resources to wealthier areas.
While planting trees can be beneficial, it must be paired with municipal maintenance strategies to prevent the decay of green spaces within struggling communities.
This research was presented at University of Kentucky's virtual Dimensions of Political Ecology conference in spring 2022, and Utah's Research on Capitol Hill in 2024:
E. P. Sargeant, M. Shcheglovitova. 2022. Does Planting Trees Improve Neighborhoods? Dimensions of Political Ecology. Virtual. March 25.E. P. Sargeant, M. Shcheglovitova. 2024. Can Algorithms and Urban Greening Repair Historic Injustice? Research on Capitol Hill, Salt Lake City, Utah. Jan 8.
E. P. Sargeant, M. Shcheglovitova. 2022. Does Planting Trees Improve Neighborhoods? Dimensions of Political Ecology. Virtual. March 25.E. P. Sargeant, M. Shcheglovitova. 2024. Can Algorithms and Urban Greening Repair Historic Injustice? Research on Capitol Hill, Salt Lake City, Utah. Jan 8.
Page updated
Google Sites
Report abuse