Find What Works: Schools

← All proof trees

PhiladelphiaAlgorithm

Reading the state's wide-format datafile

The state's Future Ready datafile has one row per school and hundreds of columns named measure_subgroup. The loader recognizes a fixed list of measure prefixes and subgroup suffixes, joins the row to a Philadelphia school through state IDs, and stores one cell per recognized column.

  1. 1
    How it was computedSteps
    1. Map AUN + Schl → ULCS via the 2024-25 Master School List; rows for schools not in the directory are dropped.
    2. For each recognized measure × subgroup column, store the value; 'IS', '--', '*' → suppressed.
    3. Store under the year the measure describes: the attendance indicator lags one year (workbook year − 1, per PDE's glossary); every other measure keeps the workbook's report year.
    4. Older long-format files (2018-19, 2019-20, 2020-21 workbooks) are read row-by-row by a separate loader (futurereadypa_longformat_attendance.ts), keeping only the 'Percent Regular Attendance' element.

    Inputs: futurereadypa_performance_<year>

    Evidence: futurereadypa_performance.ts — search "function parseColumn"

  2. 2
    Assumptions & caveatsAssumptions
    • Column names are stable across the 2021-22 to 2024-25 files (verified from first-fetch findings before loading each year).
    • Schools closed before 2024-25 are absent from every year (survivor-only panel).