Transportation routes for Baltimore City public school students are created at the beginning of each school year. However, over the course of the term, students unfortunately become homeless and need to re-locate. In order to ensure that homeless students can continue to attend the same school and maintain an element of stability during a turbulent time, the school district provides transportation accommodations for these students. These additional accommodations can often be costly, as students' new locations can frequently be outside of existing transportation routes established at the beginning of the year.
If a model was developed to help identify factors leading to homelessness and anticipate homeless students' re-locations, these changes could be accounted for during the initial route construction at the start of the school year. This would provide more efficient and reliable transportation for homeless students and savings to the school district. Additionally, the model itself could be used to better understand the driving forces behind student homelessness, which could help in mitigating its effects.
Timeline, Resources and Group Skills
The primary resource that will be needed is access to data. Receiving access to student data that identifies whether or not a student is or has been homeless, along with relevant socio-economic data, will be critical in constructing a model. Additionally, access to administrators, teachers and students will be very helpful in identifying possible factors that could indicate that students will be moving due to homelessness.
Once the initial data has been gathered, an exploratory data analysis can be performed to identify broad trends and provide the context for next steps. Since some team members already have data analysis skills and can leverage open-source software, no resources would be needed initially, and the preliminary analysis could likely be completed in 3-4 weeks. However, once the project shifts from identifying causes of homelessness to determining how homeless students re-locate, additional resources might be needed to procure software packages that specialize in analyzing geospatial data. The exact extent of the analytical work that can be done and the necessary tools that will be needed can be more accurately assessed once the initial data is received.
Hackers: Andrew Molchan, Brigitte Granger
Collaborators: Shayna Robinson, John Land, Robin Neal, Lara Ohanian