01 — Inventory (Philadelphia data landscape)
What's publicly available, in what format, at what cadence, for what years.
This is the source-of-truth catalog for the Philly build, mirroring the
implicit catalog behind verify/sources.yaml in the NYC build.
A. Publishers
| Publisher | Role | Coverage |
|---|---|---|
| School District of Philadelphia — Research, Evaluation & Accountability | District-level operational data + SPREE accountability composite | SDP district schools (all years back to 2012-13) |
| OpenDataPhilly | Republishes SDP data under stable Socrata-compatible APIs | Same as SDP |
| Pennsylvania Department of Education (PDE) | State assessment results (PSSA + Keystone), enrollment, charter school directory, statewide reporting | All Pennsylvania schools (district + charter); covers Philly schools at the same grain as the rest of the state |
| Future Ready PA Index (futurereadypa.org) | PDE's ESSA accountability dashboard; school-level downloads | All PA public schools |
| NCES (National Center for Education Statistics) | Common Core of Data (CCD) for the school directory; NAEP for state-level reference | All US public schools |
B. Sources, by analytical input
Mapping each input the NYC platform consumes to its Philly equivalent. Format "file?" means we know the publisher posts it but the exact URL needs first-fetch.
School directory + universe tagging
| Need | Source | Format | Cadence | Notes |
|---|---|---|---|---|
| All Philly schools with stable IDs | SDP + PDE + OpenDataPhilly merger | CSV (Socrata API) | Annual | SDP School ID for district; PDE 9-digit Public School Code for everything (including charters). Use PDE code as primary key (covers both sectors) and carry SDP ID as join. |
| School type (district / charter / Renaissance / cyber) | PDE charter list + SDP charter office | xlsx | Annual | Renaissance charters are a third sector (charters that converted from district schools 2010–2015); analytically meaningful, distinct from regular charters. |
| Grade band, latest enrollment, demographics | SDP "School Profiles" data + PDE "Public School Enrollment Reports" | CSV / xlsx | Annual | SDP has richer per-school demographics; PDE is the source of truth for cross-charter consistency. |
| Geography (council district, neighborhood, learning network) | SDP + OpenDataPhilly (boundaries) | GeoJSON + CSV | Static | Learning Networks (SDP's organizational sub-groupings) replace NYC's community districts as the primary geographic decomposition. |
Outcomes — test scores
| Need | Source | Format | Cadence | Notes |
|---|---|---|---|---|
| PSSA grades 3–8 ELA + Math + Science | PDE "Assessment Reporting" | xlsx + CSV | Annual (released ~December for prior year) | Subgroup breakdowns: ELL, IEP, Gender, Ethnicity, Economically Disadvantaged. Suppression for n<11. Comparability: 2014-15 PA Core redesign; 2022-23 post-COVID recalibration. |
| Keystone Algebra I / Literature / Biology | PDE "Assessment Reporting" | xlsx | Annual | End-of-course; same subgroup breakdowns. Aggregation rule for "school proficiency" is non-trivial (see O.5 in 00_overview.md). |
| Mirror copy with Philly schools only | OpenDataPhilly + SDP "PSSA Data" dataset | CSV (Socrata API) | Annual | Convenient for development; PDE is authoritative for validation. |
Outcomes — graduation, attendance, climate
| Need | Source | Format | Cadence | Notes |
|---|---|---|---|---|
| 4-year cohort graduation rate | Future Ready PA Index data files + PDE Cohort Reports | xlsx | Annual | Both 4-year and 6-year cohorts published. Subgroup breakdowns. |
| Chronic absenteeism (≥10% of enrolled days) | Future Ready PA Index | xlsx | Annual | Same definition as NYC. |
| Truancy / regular attendance | SDP "Attendance & Truancy" dataset | CSV | Annual | Finer-grained than Future Ready (per-grade, by-month at school level). |
| Climate / safety (suspensions, incidents) | SDP "School Climate" dataset + PDE Safe Schools Online | CSV | Annual | SDP publishes detailed incident counts; PDE Safe Schools is the state-comparable version. |
Climate survey
| Need | Source | Format | Cadence | Notes |
|---|---|---|---|---|
| Philly School Experience Survey (PSES) — student, parent/guardian, teacher, principal, support-staff | SDP Research dept | Zip (CSV per respondent + codebook) | Annual (since 2017-18) | 6 domains: School Climate, Instructional Environment, School Leadership, Professional Capacity, Family Engagement, DEI. More respondent groups than NYC's NYCSS (which has just student + teacher + parent). 95,000+ respondents in 2023-24. |
Accountability composites — to validate against, not invent
| Need | Source | Format | Cadence | Notes |
|---|---|---|---|---|
| SPREE (School Progress Report on Education and Equity) | SDP Performance Office | xlsx + interactive dashboard | Annual | The Philly analogue of NYC's School Quality Snapshot. Publishes per-school composite + sub-component scores. Comparable back to 2012-13. Reference for validation check (iii). |
| SPOTlight (School Performance Outcomes Table) | SDP Performance Office | xlsx | Annual | Companion to SPREE; raw measures rather than composite. |
| Future Ready PA Index dashboard | PDE | Web + xlsx | Annual | State-level accountability composite; bands schools red/yellow/green on a fixed set of indicators. |
C. Open data infrastructure
Philly has a meaningfully better open-data posture than NYC:
- OpenDataPhilly — Socrata-compatible, machine-readable, well-documented, with
provenance.json-like metadata baked in. - PDE — bulk xlsx downloads with stable URLs; less polished than Socrata but reliable.
- Future Ready PA Index — a single download bundle of all school-level indicator values; updated annually.
- No equivalent of NYC's InfoHub xlsx-via-CDN sprawl. Every source above has either a stable Socrata API or a stable PDE URL.
D. Things Philly has that NYC doesn't
- One district authority publishing data with consistent IDs end-to-end. No 32-CSD multi-source reconciliation.
- A published peer-group composite (SPREE) that's open and downloadable, going back over a decade — gives us a built-in spot-check reference at validation level (iii) on day one.
- A richer climate survey (PSES) — same domains as NYC plus DEI; five respondent groups instead of three.
- Better state-level accountability data (Future Ready PA Index) — NYSED's analogue is messier.
E. Things NYC has that Philly doesn't
- A bigger civic-data ecosystem — Chalkbeat NY, NY Focus, The 74, InsideSchools, NYU Steinhardt, etc., generate prior reporting and named experts to interview. Philly has Chalkbeat Philadelphia and a thinner academic press.
- Detailed per-school admission category at NYC's granularity (SCREENED / OPEN / ZONED / AUDITION / G&T / SPECIALIZED HS / etc.). PA classifies more simply (district-zoned vs charter-lottery vs specialized magnet); the taxonomy will be smaller.
- The InsideSchools narrative layer — qualitative descriptions per school. Philly's Great Philly Schools (greatphillyschools.org) is the closest analogue but smaller in coverage.
F. Things both publish at comparable quality
- Test scores (school × grade × subject × subgroup × year)
- Graduation rates
- Chronic absenteeism
- Demographic shares per school per year
- Survey data
G. Things that genuinely require new work
- Cyber-charter handling. Pennsylvania's cyber-charter sector is the largest in the country; many serve Philly students. v1 excludes them; v2 needs a tagged-subset treatment.
- Renaissance Charter sector. No NYC analogue. Need a third school_type enum value or a separate flag.
- Council district + neighborhood geography. OpenDataPhilly publishes the boundaries; we'd do the point-in-polygon join. Not done in NYC (we used borough + community district).
- Statewide reference cohorts. "Philly schools vs. similar-demographic Pennsylvania schools" is a story type that doesn't exist in NYC; would need PA-wide ingest if pursued (out of scope for v1).