In 2015, the Bluff City Snapshot (BCS) was launched to document the condition of every parcel in Memphis. Over 12 weeks, more than 100 dedicated volunteers traversed the city, surveying properties to gather critical data on their condition. From structural integrity to the presence of litter, the BCS aimed to provide a comprehensive view of the state of properties across the city, which could be used to inform urban planning and policy-making efforts.
The Bluff City Snapshot survey involved rating properties on a scale of 1 to 5, with 1 indicating a property in excellent condition and 5 representing severe blight. This rating was based on a range of factors, including the condition of the structure, presence of overgrown vegetation, and whether the property appeared vacant or occupied. Each volunteer surveyor was trained to assess these variables, but some of the evaluation relied on subjective judgment.
This subjectivity has led to an important question: can we reliably track changes in a property’s condition over time using this data? The Center for Applied Earth Science and Engineering Research (CAESER) was tasked with evaluating whether the data collected in the survey could accurately explain the overall blight score assigned to each property.
After analyzing data from over 230,000 properties, we found a clear pattern: the more issues identified in a property (such as litter, structural damage, or fire damage), the higher the blight score. However, while this relationship was evident, the accuracy of the blight score as a long-term indicator of property condition remains uncertain.
One significant finding was that the survey data could only predict a property’s blight score with about 50% accuracy. This suggests that while the blight score can offer a general sense of property condition, it may not be reliable enough to track a property’s progress over time. More work is needed to refine the scoring system or supplement the survey with additional data that could improve its accuracy.
Despite the limitations, the Bluff City Snapshot remains a valuable tool for understanding the overall state of Memphis neighborhoods. It provides city officials, urban planners, and community organizations with a bird’s-eye view of areas that may require intervention, from blight reduction initiatives to redevelopment projects. The data collected helps to inform policies that aim to improve living conditions and promote community well-being.
The next step in this work could be developing a more rigorous and consistent measure of property condition that minimizes subjectivity and offers a clearer picture of long-term changes in neighborhood health. This would make it easier to track the progress of urban renewal efforts and better allocate resources to the areas that need them most.
For more technical readers, the analysis used a dataset consisting primarily of boolean variables, indicating the presence or absence of specific conditions (such as litter or fire damage). The blight score, an ordinal value from 1 to 5, was the target variable. After cleaning the data and removing records with missing values, a total of 21 features remained for analysis.
The key method used to explore the relationship between the variables and the blight score was a random forest model, which was chosen due to its ability to handle complex interactions between variables. While the model did reveal some patterns, its predictive accuracy was limited, suggesting that additional variables or more precise rating criteria may be needed to fully explain the blight score.
