Datasets:

rasinmuhammed
/

misata-edtech-practice

attempt_id int64 23 10M	student_id int64 190 100k	attempted_at stringlengths 29 29	correct bool 2 classes	duration_seconds int64 1 1.58k
3,438,887	85,891	2025-01-01 00:03:02.896483891	true	633
6,173,357	15,414	2025-01-01 00:14:50.902080670	false	214
8,396,886	19,472	2025-01-01 00:32:43.802296546	false	435
949,189	53,763	2025-01-01 00:35:50.165894510	false	250
6,295,692	19,463	2025-01-01 01:12:16.438461583	false	108
4,125,344	71,096	2025-01-01 01:50:40.733098036	false	150
3,597,579	46,123	2025-01-01 02:13:07.836177565	false	77
2,480,580	53,776	2025-01-01 02:25:49.482188995	false	69
9,698,898	94,470	2025-01-01 02:40:51.207011500	true	31
3,173,648	69,708	2025-01-01 02:42:59.622182718	false	62
7,482,403	16,523	2025-01-01 03:16:19.658965927	false	68
880,389	93,502	2025-01-01 03:33:10.191974817	false	103
5,737,855	22,548	2025-01-01 04:40:02.142698605	true	768
8,663,377	32,830	2025-01-01 05:25:19.355313887	true	287
7,923,672	22,548	2025-01-01 05:37:43.583576722	false	214
1,278,515	7,822	2025-01-01 05:42:27.753599565	true	226
6,335,919	14,285	2025-01-01 06:27:15.333169158	true	440
8,104,429	37,637	2025-01-01 06:41:01.126997050	false	293
9,767,641	27,731	2025-01-01 06:41:29.892993757	false	260
2,932,374	82,912	2025-01-01 07:21:09.115792416	false	174
8,397,218	86,534	2025-01-01 07:36:38.742880919	false	83
5,309,253	61,943	2025-01-01 07:51:37.080381742	true	78
2,986,640	22,192	2025-01-01 08:47:29.433257986	false	178
6,941,930	5,413	2025-01-01 09:03:05.261658766	true	281
1,788,288	18,298	2025-01-01 09:08:22.402121023	true	97
7,575,377	72,890	2025-01-01 09:51:11.470718158	false	215
8,922,381	24,238	2025-01-01 10:02:44.190859987	false	132
9,351,639	47,784	2025-01-01 10:41:09.476345605	false	297
4,560,373	52,415	2025-01-01 10:58:41.757919195	false	254
4,741,970	41,148	2025-01-01 11:00:23.934338381	false	45
3,937,891	98,442	2025-01-01 11:03:18.897588201	false	563
7,729,118	84,141	2025-01-01 11:16:02.687255197	true	527
7,652,397	45,774	2025-01-01 11:42:30.214325672	true	84
167,396	34,205	2025-01-01 12:02:52.958208595	true	68
5,218,458	20,429	2025-01-01 12:22:40.535028913	true	234
9,166,941	52,610	2025-01-01 12:35:59.961159590	true	194
900,097	41,882	2025-01-01 12:42:09.341688759	false	350
3,995,747	85,359	2025-01-01 12:48:24.615218802	true	128
3,539,590	34,205	2025-01-01 12:53:46.826866956	false	219
4,540,126	65,369	2025-01-01 13:01:44.728092899	true	408
238,551	52,675	2025-01-01 13:58:01.921809679	false	314
2,782,102	97,474	2025-01-01 14:03:10.305124008	false	359
5,106,783	48,806	2025-01-01 14:37:07.616313881	false	107
4,244,586	25,921	2025-01-01 14:47:12.225732320	true	188
4,820,668	53,070	2025-01-01 15:03:15.435673905	false	276
2,335,097	45,537	2025-01-01 15:06:42.082631619	false	497
2,071,267	33,866	2025-01-01 15:10:19.890241904	false	124
3,031,200	93,217	2025-01-01 15:13:39.896923027	true	293
5,899,994	91,932	2025-01-01 15:34:33.243161230	true	49
9,366,067	49,847	2025-01-01 15:35:27.182231346	true	72
5,971,639	33,866	2025-01-01 16:08:16.744466170	false	170
4,059,443	29,756	2025-01-01 16:20:18.900443043	true	481
5,327,723	8,854	2025-01-01 16:24:12.068783247	false	251
863,571	72,807	2025-01-01 16:47:05.867369722	true	171
6,583,576	31,894	2025-01-01 17:11:55.387068749	false	216
1,265,468	48,768	2025-01-01 17:12:35.776783535	true	26
4,114,855	26,735	2025-01-01 17:44:09.761286535	true	736
9,830,956	71,836	2025-01-01 17:53:25.555133878	true	77
5,437,753	53,070	2025-01-01 18:29:16.742548544	true	32
5,546,481	72,677	2025-01-01 18:38:25.234275783	true	200
8,856,133	63,291	2025-01-01 18:43:36.665552097	false	89
6,314,416	62,846	2025-01-01 19:29:49.776572768	true	102
9,496,407	190	2025-01-01 19:38:17.063918857	true	70
8,030,321	67,311	2025-01-01 19:58:27.020980607	true	391
8,805,315	88,209	2025-01-01 20:00:29.772018377	true	72
2,575,971	19,155	2025-01-01 20:01:59.664086781	true	141
7,435,014	95,836	2025-01-01 20:03:29.528105178	true	73
4,312,690	24,050	2025-01-01 20:28:50.725172129	true	308
3,746,532	56,818	2025-01-01 20:29:52.881066657	false	105
5,957,857	68,090	2025-01-01 20:53:50.497240855	true	182
8,582,047	24,380	2025-01-01 21:02:33.849256082	true	265
1,799,154	89,170	2025-01-01 21:20:29.259181407	true	278
2,081,665	27,132	2025-01-01 21:46:00.817837704	true	75
8,932,029	81,888	2025-01-01 22:01:16.117698735	true	242
5,875,880	22,192	2025-01-01 22:54:24.992598977	false	253
2,932,796	50,913	2025-01-01 22:57:37.072999456	true	164
8,484,598	35,379	2025-01-01 23:06:56.522875041	true	406
1,410,230	65,666	2025-01-01 23:15:25.751019915	true	62
4,789,691	91,932	2025-01-01 23:56:37.333309418	true	122
2,134,751	53,047	2025-01-02 00:28:56.340984189	true	118
9,745,179	98,410	2025-01-02 00:42:35.767134291	true	358
7,964,922	57,204	2025-01-02 01:22:21.427829865	false	184
1,774,328	77,195	2025-01-02 01:23:40.207222765	true	408
1,506,213	5,965	2025-01-02 02:03:57.300729126	true	92
7,514,087	98,006	2025-01-02 02:23:15.631288407	true	294
423,365	53,763	2025-01-02 02:25:31.439013967	true	441
4,750,975	687	2025-01-02 02:47:24.515993818	true	137
6,537,307	20,429	2025-01-02 02:52:55.489808921	true	471
3,774,313	86,534	2025-01-02 04:03:19.415141457	false	190
3,180,278	66,380	2025-01-02 04:13:02.123367596	false	218
2,757,645	1,368	2025-01-02 04:25:09.590341337	false	77
245,717	39,436	2025-01-02 05:14:06.854381786	true	180
8,381,706	23,527	2025-01-02 05:15:03.395254657	false	327
9,841,204	56,207	2025-01-02 05:28:30.826791835	true	81
6,996,735	87,503	2025-01-02 05:29:18.758137799	false	127
3,751,516	88,068	2025-01-02 05:53:05.767643637	true	97
730,135	80,082	2025-01-02 06:22:26.745369371	false	254
5,845,510	52,922	2025-01-02 06:31:52.935622700	false	107
604,217	23,487	2025-01-02 07:12:47.501049424	true	256
2,943,329	22,085	2025-01-02 07:28:15.269705552	true	69

End of preview. Expand in Data Studio

Misata EdTech Practice Evalpack (declared learning curve)

An evalpack: an evaluation database generated from the answer key, not annotated after the fact. A VLDB 2026 audit found 52.8% of BIRD Mini-Dev answer keys wrong because benchmarks annotate answers onto existing databases; this dataset inverts the order. The declared properties (curves, shares, identities) are the specification, the database is generated to satisfy them exactly, and every shipped question was re-verified by executing its gold SQL against these exact files with DuckDB, an engine that shares no code with the generator.

What is in it

Tables: students (600 rows), practice_attempts (15,000 rows).

A practice log where the monthly correct rate climbs 55% to 68% by declaration and total practice time per month is declared and exact. Every attempt belongs to a student who exists and happens after that student signed up. 17 questions ship; 4 candidates whose rates were unachievable at the generated row counts were dropped by the verification gate rather than shipped wrong (that refusal is the point).

Files

tables/*.csv: the database
questions.jsonl: one verified question per line (natural language, gold SQL, expected answer, tags)
certificate.json: per-question DuckDB verification plus FK proof
manifest.json: spec hash, seed, library versions, dropped candidates
verify.py: standalone re-verification (needs only pip install duckdb)
schema.misata.yaml: the full declaration; regenerate or rotate the pack from it

Re-verify in thirty seconds

pip install duckdb
python verify.py

Rotate the environment without touching the answer key

pip install misata
misata evalpack --config schema.misata.yaml -o rotated_pack --seed 7

A new seed replaces the rows; the declared answers stay the answers, so agents cannot memorize the environment. Same version + same schema + same seed reproduces these bytes exactly.

Provenance

No real data: fully declaration-generated, offline, deterministic (provenance statement). Generated with misata (MIT). Where the approach fails is documented in LIMITATIONS.md.

Downloads last month: 40