Dataset Demand List and Priority (#530)¶
Ranked list of datasets to convert from catalogue to spec next. Every entry has a reason. Nothing without a reason is on this list.
Relation to docs/dataset-prioritization.yaml. That file scores datasets on
demand, pain, openness and stability and is the long-term ranking. This list is the
work queue: it puts first what builder is blocked on and what users asked for, so
its order can differ from the scores. When they disagree, this list decides what is
converted next; the YAML decides where a dataset lands once the queue is empty.
Tier 1 — Builder already uses these (blocked on spec conversion)¶
These datasets are in builder's pipeline configs. Rows 1–10 are already spec; the DUR family (rows 11–19) is still catalogue-only in kpubdata. Converting it unlocks builder's #636 (rebuild HF datasets through the engine).
| # | Dataset | Reason | Builder Config |
|---|---|---|---|
| 1 | datago.apt_trade |
Paper dataset. Builder param_grid 1,500 combos. Already spec ✅ | seoul_apartment_trades |
| 2 | datago.apt_rent |
Paper dataset. Builder param_grid 1,500 combos. Already spec ✅ | seoul_apartment_rent |
| 3 | bok.base_rate |
Builder uses for interest rate analysis. No activation needed. | korea_base_rate |
| 4 | datago.ultra_srt_ncst |
Builder weather pipeline. Already spec ✅ | weather_ultra_srt_ncst |
| 5 | datago.village_fcst |
Builder weather pipeline. Already spec ✅ | weather_village_fcst |
| 6 | datago.air_quality |
Builder pipeline. Already spec ✅ | air_quality |
| 7 | datago.tour_kor_area |
Builder tourism pipeline. Already spec ✅ | tour_kor_area |
| 8 | datago.tour_kor_festival |
Builder tourism pipeline. Already spec ✅ | tour_kor_festival |
| 9 | datago.tour_kor_keyword |
Builder tourism pipeline. Already spec ✅ | tour_kor_keyword |
| 10 | datago.tour_kor_location |
Builder tourism pipeline. Already spec ✅ | tour_kor_location |
| 11–19 | datago.dur_* (9) |
DUR (Drug-Use Review) family. Builder has 9 pipeline configs. All are data.go.kr JSON with standard envelope. | dur_* |
Tier 2 — Requested by users (dataset-request issues)¶
| # | Dataset | Reason | Issue |
|---|---|---|---|
| 20 | datago.offi_trade_view |
Officetel sale — Dev variant requested | #402 |
| 21 | datago.apt_rent_view |
Apt rent Dev variant requested | #401 |
| 22 | datago.school_info |
NEIS school basic info via data.go.kr | #400 |
| 23 | datago.airkorea_forecast |
Already spec ✅ as airkorea_forecast (requested as air_forecast) |
#399 |
| 24 | datago.ultra_srt_fcst |
Already spec ✅ (was #398) | #398 |
| 25 | datago.air_station |
Already spec ✅ (was #396) | #396 |
Tier 3 — High-value catalogue datasets (high volume, public interest)¶
| # | Dataset | Reason |
|---|---|---|
| 26 | localdata.* (59) |
Marked retired by #603; the evidence disagrees (3 answered, 56 need activation — #618). Do not convert until #618 is decided. |
| 27 | sgis.boundary.* (3) |
Spatial analysis foundation. Token auth is unique. |
| 28 | krx.kospi_index |
Financial analysis. No auth needed. |
| 29 | kosis.population_migration |
Demographic analysis. Already requested for paper. |
Already Converted (23 spec datasets — no action needed)¶
All 23 spec datasets are listed in SUPPORTED_DATA.md. 22 pass make verify with
recorded fixtures; datago.ocean_buoy has no fixture yet and is listed as 진행 중
(spec status: unstable).
Summary¶
- Already spec: 23 datasets (real estate 7, weather 3, air 3, tourism 4, hospital, metro fare, ocean buoy, localdata 3)
- Next priority: 9 DUR datasets (builder uses them, standard envelope, mechanical conversion)
- After that: 3 requested datasets from issues #400–#402 (#399 is done as
airkorea_forecast) - On hold: 59 localdata datasets until #618 settles whether they are retired