UDP - Student Mobility

1. 'base_files': Import and clean raw data, combine to create analytical files

tps_covariates.py

Import 'student_geos_.csv' (base) files (provided by TPS)

Aggregate files together, reshape (long by tract, wide by year)

Export 'tps_student_counts_by_tract.csv'

tulsa_parcels.py

Import 'TulsaParcelsNeighborhood.csv' (base) file (provided by CoT)

Keep desired fields

Export 'tulsa_parcels.csv'

tulsa_parcel_tps_tract_association.Rmd

Import 'tulsa_parcels.csv'

Associate parcel geocodes to geotracts (same as that in 'tps_student_counts_by_tract.csv'

Export 'tulsa_parcel_tps_tract_association.csv'

---

base_analytical_file.py

Import 'tps_student_counts_by_tract.csv' (Student Count Data)

Import 'tulsa_parcel_tps_tract_association.csv' (Parcel Data)

Filter, normalize, reshape, split (by parcel types)

Merge together parcel and student count data

Export 'base_file_‹wide/long›_‹residential/commercial/combo›.csv' (analytical) files

2. 'analysis': Run Analysis on analytical files

feature_selection_methods.py

Define feature selection methods

udp_feature_selection.py

Iterate over each group type (residential, commercial, combo) and year (2013-2018) combination:

Apply each feature selection method to the target year, using parcel data from prior years

Combine year-to-year results for each group type

Export 'analysis_‹residential/commercial/combo›.csv' (analytical result) files