DAAF-Contribution-Community/daaf / .claude/skills/education-data-explorer
education-data-explorer
VerifiedMirror-first discovery for Urban Institute Education Data: maps a research question to the mirror dataset FILES that answer it — canonical path, entity grain, year coverage, key variables, and join keys — for CCD, IPEDS, CRDC, College Scorecard, SAIPE, EDFacts, FSA, MEPS, NHGIS, PSEO, NACUBO, NCCS, EADA, and Campus Safety. Discovery output is the mirror path (plus grain, years, variables, caveats), the currency planning and acquisition consume. A pure readable-reference skill for read-only discovery agents — no code, no live HTTP probes; all facts keyed to pinned mirror vintage 0.26.1. Load BEFORE education-data-query — routing (which file, what grain) here, retrieval (fetch code, mirror config) there.
references/districts-datasets.md
View on GitHub# Districts Datasets Reference (mirror-first)
District/LEA-level mirror dataset families for the Urban Institute Education Data mirror. Each entry gives the **canonical mirror path** (the acquisition currency — extensionless `{source}/{filename}`, with `{year}` where the family is Yearly), the entity **grain**, **year coverage** (vintage 0.26.1), curated **key variables** (real names — full lists in `variable-dictionary-{source}.md`), **join keys**, **caveats**, and a small **Portal endpoint** annotation for citation/maintenance only.
> **Vintage:** all facts keyed to pinned mirror vintage **0.26.1** (revision `0ad00ce0e232c96b0642459e4e7326607a8d26aa`). The `urban_csv` fallback is unpinned/current-Portal. Primary ID at this level is `leaid` (7-char zero-padded string after normalization; raw mirror files store it as String or as Int64 without leading zeros, depending on the file). Emit paths, not routes. Do not instruct fetching or probing — retrieval belongs to `education-data-query`.
---
## CCD (Common Core of Data)
### CCD District Directory & Staffing — `ccd/school-districts_lea_directory`
- **Type / years:** Single-file · 1986-2024
- **Grain:** one row per `leaid` × `year` (LEA directory + staffing counts).
- **Key variables:** `lea_name`, `fips`, `enrollment`, `number_of_schools`, `urban_centric_locale`, `agency_type`, `agency_level`, `lowest_grade_offered`, `highest_grade_offered`, `teachers_total_fte`, `spec_ed_students`, `english_language_learners` (full 69-var list: `variable-dictionary-ccd.md` § `ccd/school-districts_lea_directory`).
- **Join keys:** `leaid` (+ `year`) to any district source; schools roll up to their LEA via `leaid`.
- **Caveats:** `agency_type` is **not** a Portal filter despite older docs (use post-fetch filtering). FTE staffing counts are LEA-reported and can carry `-1`/`-2` sentinels.
- **Codebook:** `ccd/codebook_districts_ccd_directory`
- **Portal endpoint(s):** `/school-districts/ccd/directory/{year}/`
### CCD District Finance — `ccd/districts_ccd_finance`
- **Type / years:** Single-file · 1989, 1991, 1994-2020 (non-contiguous early years; the mirror file holds 15,931 rows for 1989, a year the Portal catalog does not list, as it shows 1991, 1994-2020)
- **Grain:** one row per `leaid` × `year` (F-33 fiscal record).
- **Key variables:** `rev_total`, `rev_fed_total`, `rev_state_total`, `rev_local_total`, `rev_local_prop_tax`, `exp_total`, `exp_current_elsec_total`, `exp_current_instruction_total`, `exp_current_supp_serve_total`, `salaries_total`, `salaries_instruction`, `enrollment_fall_responsible` (full 163-var list — includes CARES/CRRSA/ARP ESSER relief columns and detailed rev/exp/salary/benefit/debt breakdowns: `variable-dictionary-ccd.md` § `ccd/districts_ccd_finance`).
- **Per-pupil:** compute from `exp_current_elsec_total` / enrollment; the mirror finance file stores dollar totals, not pre-divided per-pupil columns.
- **Join keys:** `leaid` (+ `year`) — commonly joined to `ccd/school-districts_lea_directory` (characteristics) and `saipe/districts_saipe` (poverty). This file stores `leaid` as String while those two store Int64 (pinned mirror, probed 2026-10-28), so name the cast-to-String + pad-to-7 normalization as a join caveat.
- **Caveats:** fiscal year lags the school year (~2 years); mirror coverage ends 2020. Census `censusid` present for cross-linking to Census fiscal data.
- **Codebook:** `ccd/codebook_districts_ccd_finance`
- **Portal endpoint(s):** `/school-districts/ccd/finance/{year}/`
### CCD District Enrollment — `ccd/schools_ccd_lea_enrollment_{year}`
- **Type / years:** Yearly · 1986-2024 (one file per year)
- **Grain:** one row per `leaid` × `year` × `grade` × `race` × `sex`. Disaggregation dimensions are **columns**, not URL segments — filter to the `99`/total code on each dimension for an LEA total (do not sum, or totals rows double-count).
- **Key variables:** `leaid`, `year`, `fips`, `grade`, `race`, `sex`, `enrollment` (complete 7-var family: `variable-dictionary-ccd.md` § `ccd/schools_ccd_lea_enrollment_{year}`).
- **Join keys:** `leaid` (+ `year`, and dimension codes where relevant).
- **Caveats:** `grade=-1` is Pre-K (not missing); `grade=99` is all-grades total. Fetch each needed year and concatenate.
- **Codebook:** `ccd/codebook_districts_ccd_enrollment`
- **Portal endpoint(s):** `/school-districts/ccd/enrollment/{year}/{grade}/` (+ `/race/`, `/sex/`, `/race/sex/`)
---
## SAIPE (Small Area Income and Poverty Estimates)
### District Poverty — `saipe/districts_saipe`
- **Type / years:** Single-file · 1995, 1997, 1999-2024 (non-contiguous early years)
- **Grain:** one row per `leaid` × `year` (Census model-based district poverty estimate).
- **Key variables:** `district_id`, `district_name`, `est_population_total`, `est_population_5_17`, `est_population_5_17_pct`, `est_population_5_17_poverty`, `est_population_5_17_poverty_pct` (complete 10-var family: `variable-dictionary-saipe.md`).
- **Join keys:** `leaid` (+ `year`) to CCD district directory/finance and EDFacts.
- **Caveats:** `est_population_5_17_poverty_pct` is on a **0-1 scale despite the `_pct` suffix** — verify scale before treating as a percentage. No race/ethnicity disaggregation at district level (use ACS 5-year for that). Released ~12-13 months after the income year ends (DAAF inference from Census release dates; 2024 estimates released Jan 27, 2026). R readers: this file has no `string_view` columns at the pinned revision (probed 2026-10-28), but use the view-safe read from `education-data-query` for every mirror file, because 75 others do.
- **Codebook:** `saipe/codebook_districts_saipe`
- **Portal endpoint(s):** `/school-districts/saipe/{year}/`
---
## EDFacts
### District Assessments — `edfacts/districts_edfacts_assessments_{year}`
- **Type / years:** Yearly · 2009-2018, 2020 (**2019 absent** — COVID testing waivers)
- **Grain:** one row per `leaid` × `year` × `grade_edfacts` × subgroup dimension. Subgroups are columns (`race`, `sex`, `lep`, `disability`, `econ_disadvantaged`, `homeless`, `migrant`, `foster_care`, `military_connected`).
- **Key variables:** `grade_edfacts`, `read_test_num_valid`, `read_test_pct_prof_midpt` (+ `_low`/`_high`), `math_test_num_valid`, `math_test_pct_prof_midpt` (+ `_low`/`_high`) (full 23-var list: `variable-dictionary-edfacts.md` § `edfacts/districts_edfacts_assessments_{year}`).
- **Join keys:** `leaid` (+ `year`, `grade_edfacts`, subgroup codes).
- **Caveats:** proficiency is **range-reported** (`_low`/`_midpt`/`_high`) for small-cell protection — use the midpoint and carry the range. `grade_edfacts` codes are `{3-8, 9="Grades 9-12", 99}` (distinct from CCD numeric `grade`). **State assessment scores cannot be compared across states** (different tests/cut scores). `disability` uses the full 0-4/99 scheme (see `variable-codes.md`).
- **Codebook:** `edfacts/codebook_districts_edfacts_assessments`
- **Portal endpoint(s):** `/school-districts/edfacts/assessments/{year}/{grade_edfacts}/` (+ `/race/`, `/sex/`, `/special-populations/`)
### District Graduation Rates — `edfacts/districts_edfacts_grad_rates_{year}`
- **Type / years:** Yearly · 2010-2019
- **Grain:** one row per `leaid` × `year` × subgroup (ACGR — adjusted cohort graduation rate).
- **Key variables:** `cohort_num`, `grad_rate_midpt` (+ `_low`/`_high`), plus subgroup columns `race`, `lep`, `disability`, `econ_disadvantaged`, `homeless`, `foster_care` (full 15-var list: `variable-dictionary-edfacts.md` § `edfacts/districts_edfacts_grad_rates_{year}`).
- **Join keys:** `leaid` (+ `year`, subgroup codes).
- **Caveats:** grad rate is **range-reported** (use `grad_rate_midpt`). ACGR is a 4-year adjusted cohort measure; not comparable to CCD/IPEDS completion counts.
- **Codebook:** `edfacts/codebook_districts_edfacts_graduation`
- **Portal endpoint(s):** `/school-districts/edfacts/grad-rates/{year}/`
---
## Common District Use Cases
| Question | Files (paths) | Join |
|----------|---------------|------|
| District characteristics + spending | `ccd/school-districts_lea_directory` + `ccd/districts_ccd_finance` | `leaid` + `year` |
| Poverty and achievement | `saipe/districts_saipe` + `edfacts/districts_edfacts_assessments_{year}` | `leaid` + `year` |
| Enrollment trends | `ccd/schools_ccd_lea_enrollment_{year}` (multi-year) | `leaid` + `year` (+ dimension totals) |
| Spending and graduation | `ccd/districts_ccd_finance` + `edfacts/districts_edfacts_grad_rates_{year}` | `leaid` + `year` |
Related skills
| Skill | Stars | Updated | Works with | Status | Install |
|---|---|---|---|---|---|
| finishing-a-development-branchobra/superpowersUse when implementation is complete, all tests pass, and you need to decide how to integrate the work | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |
| systematic-debuggingobra/superpowersUse when encountering any bug, test failure, or unexpected behavior, before proposing fixes | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |
| test-driven-developmentobra/superpowersUse when implementing any feature or bugfix, before writing implementation code | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |
| skill-creatoranthropics/skillsCreate new skills, modify and improve existing skills, and measure skill performance. Use when users want to create a skill from scratch, edit, or optimize an existing skill, run evals to test a skill, benchmark skill performance with variance analysis, or optimize a skill's description for better triggering accuracy. | Stars: 198k | Updated | Claude CodeCodex | Verified | Install |