Datasets:

hoodsy
/

sn-triage-dataset

date date32	site large_stringclasses 32 values	actual_kwh float64 -515.66 242k	expected_kwh float64 -223.1 237k ⌀	pi float64 -2,178,908.51 251k ⌀	pi_baseline float64 0 1.89 ⌀	coverage float64 0 1.04	rain_mm float64 0 110 ⌀	snow_cm float64 0 38.4 ⌀	flagged bool 2 classes	pi_rel float64 -2,028,542.5 10 ⌀	csr float64 -79.26 717 ⌀	model_csr float64 -0.01 1.74k ⌀	chop float64 0 40.3	plateau_run_h float64 0 4.25	peak_frac float64 -0.76 491	expected_peak_frac float64 -0 1.6 ⌀	dead_run_h float64 0 11.5	morning_ratio float64 -53.93 3.31 ⌀	midday_ratio float64 -18.46 3.43 ⌀	tmax_c float64 -24.37 50 ⌀	temp_corr float64 -1 1 ⌀	snow_3d float64 0 64.8 ⌀	snow_7d float64 0 66.7 ⌀	snow_30d float64 0 90.7 ⌀	pi_slope_14d float64 -62,254.6 62.3k ⌀	pi_drop_share float64 0.25 283M ⌀	unit_count float64 -5.59 27.9 ⌀	unit_dist float64 0 0.5 ⌀	doy_sin float64 -1 1	doy_cos float64 -1 1	rule_label large_stringclasses 10 values	final_label large_stringclasses 11 values	final_label_2 large_stringclasses 1 value	verdict large_stringclasses 4 values	provenance large_stringclasses 2 values	evidence large_stringlengths 15 253 ⌀
2010-08-23	34	151.225	null	null	null	0	null	null	true	null	0.295395	0	0	1	0.832123	null	0	null	null	null	null	null	null	null	null	null	null	null	-0.783934	-0.620844	data_gap	data_gap	null	null	rule	only 0% of intervals judgeable — meter or sensor gap, performance unknown
2010-08-24	34	34.9375	507.292296	null	null	0.291667	null	null	true	null	0.039571	0.574574	0.000375	0	0.436642	1.146642	0.5	null	-0.001419	44.172222	null	null	null	null	null	null	null	null	-0.794497	-0.607268	data_gap	data_gap	null	null	rule	only 29% of intervals judgeable — meter or sensor gap, performance unknown
2010-08-25	34	0	1,029.814036	null	null	0.447917	null	null	true	null	0	1.170324	0	0	0	1.165227	0	null	null	43.103889	null	null	null	null	null	null	null	null	-0.804826	-0.593511	data_gap	data_gap	null	null	rule	only 45% of intervals judgeable — meter or sensor gap, performance unknown
2010-08-26	34	0	910.585804	null	null	0.447917	null	null	true	null	0	1.03839	0	0	0	1.119273	0	null	null	42.777778	null	null	null	null	null	null	null	null	-0.814916	-0.579579	data_gap	data_gap	null	null	rule	only 45% of intervals judgeable — meter or sensor gap, performance unknown
2010-08-27	34	244.0125	1,017.418323	null	null	0.645833	null	null	true	null	0.27924	1.1643	0.034155	0.5	0.771186	1.143868	0	null	0.75496	40.611667	0.26148	null	null	null	null	null	null	null	-0.824765	-0.565476	data_gap	data_gap	null	null	rule	only 65% of intervals judgeable — meter or sensor gap, performance unknown
2010-08-28	34	896.475	1,118.774512	0.787134	null	0.927083	null	null	false	null	1.029587	1.284895	0.024203	1.75	0.933818	1.203692	0	0.841885	0.778545	36.044444	-0.620361	null	null	null	null	null	null	null	-0.83437	-0.551205	healthy	healthy	null	null	rule	null
2010-08-29	34	867.6125	1,073.619481	0.801716	null	0.96875	null	null	false	null	1.000102	1.237568	0.02681	1.75	0.933818	1.175491	0	0.873415	0.791158	31.184444	-0.655595	null	null	null	null	null	null	null	-0.843728	-0.536771	healthy	healthy	null	null	rule	null
2010-08-30	34	852.9	1,087.802326	0.770384	null	0.916667	null	null	false	null	0.986834	1.258624	0.022867	2	0.871671	1.162823	0	0.853689	0.740913	34.373889	-0.813689	null	null	null	null	null	null	null	-0.852837	-0.522178	healthy	healthy	null	null	rule	null
2010-08-31	34	823.05	1,086.602429	0.751977	null	0.96875	null	null	false	null	0.955949	1.262058	0.024096	2	0.842615	1.154872	0	0.838134	0.724393	35.262778	-0.916604	null	null	null	null	null	null	null	-0.861693	-0.50743	healthy	healthy	null	null	rule	null
2010-09-01	34	807.6	1,084.391479	0.739539	null	0.989583	null	null	false	null	0.941679	1.264424	0.025463	2.25	0.824859	1.160881	0	0.826299	0.709139	38.345	-0.93044	null	null	null	null	null	null	null	-0.870294	-0.492533	healthy	healthy	null	null	rule	null
2010-09-02	34	792.35	1,085.032795	0.723273	null	0.989583	null	null	false	null	0.927592	1.270231	0.029946	2	0.80791	1.162176	0	0.809649	0.694163	41.556667	-0.931649	null	null	null	null	null	null	null	-0.878637	-0.477489	healthy	healthy	null	null	rule	null
2010-09-03	34	780.65	1,083.317283	0.713826	null	0.989583	null	null	false	null	0.917625	1.273399	0.027872	2.5	0.784504	1.159494	0	0.803408	0.679059	42.26	-0.916811	null	null	null	null	null	null	null	-0.886721	-0.462305	healthy	healthy	null	null	rule	null
2010-09-04	34	797.7	1,087.742438	0.72701	null	0.864583	null	null	false	null	0.94157	1.283924	0.034187	0.75	0.820016	1.161806	0	0.798887	0.712388	41.551111	-0.860132	null	null	null	null	null	null	null	-0.894542	-0.446983	healthy	healthy	null	null	rule	null
2010-09-05	34	834.075	1,077.872891	0.757835	null	0.979167	null	null	false	null	0.988687	1.277677	0.021417	1.75	0.876513	1.164765	0	0.824841	0.7448	39.711667	-0.609435	null	null	null	null	null	null	null	-0.902099	-0.43153	healthy	healthy	null	null	rule	null
2010-09-06	34	809.3	1,070.134141	0.749626	null	0.989583	null	null	false	null	0.963473	1.273997	0.014801	1.75	0.856336	1.15293	0	0.824961	0.726631	38.063889	-0.953431	null	null	null	null	null	null	null	-0.909388	-0.415948	healthy	healthy	null	null	rule	null
2010-09-07	34	587.45	758.231771	0.774763	null	1	null	null	false	null	0.702447	0.90666	0.083394	0.25	1.079096	1.188528	0	0.874848	0.77604	32.500556	-0.301779	null	null	null	-0.006014	0.835345	null	null	-0.916409	-0.400244	healthy	healthy	null	null	rule	null
2010-09-08	34	839.45	1,063.022688	0.783238	null	0.916667	null	null	false	null	1.008293	1.276834	0.022844	1.75	0.891848	1.148954	0	0.83831	0.773253	35.622222	-0.811746	null	null	null	-0.003486	2.532809	null	null	-0.923158	-0.384421	healthy	healthy	null	null	rule	null
2010-09-09	34	817.4	1,042.43015	0.77751	null	0.989583	null	null	false	null	0.986307	1.257837	0.025989	1.75	0.859564	1.125006	0	0.849171	0.758682	31.595	-0.920445	null	null	null	-0.001568	8.043416	null	null	-0.929634	-0.368484	healthy	healthy	null	null	rule	null
2010-09-10	34	808.525	1,055.232467	0.759927	null	0.9375	null	null	false	null	0.98015	1.279226	0.02376	2	0.85795	1.145995	0	0.844064	0.73793	32.99	-0.913937	null	null	null	-0.000546	3.255801	null	null	-0.935835	-0.352439	healthy	healthy	null	null	rule	null
2010-09-11	34	799.4	1,048.458931	0.748289	null	0.927083	null	null	false	null	0.973692	1.277053	0.023848	2.25	0.838579	1.154964	0	0.830597	0.725953	35.161111	-0.905157	null	null	null	-0.00039	1.15165	null	null	-0.941759	-0.336289	healthy	healthy	null	null	rule	null
2010-09-12	34	784.375	1,028.487736	0.748769	0.757835	0.927083	null	null	false	0.988037	0.960011	1.258785	0.028581	2.25	0.820823	1.140355	0	0.830939	0.719458	36.663333	-0.835317	null	null	null	0.000268	0.586448	null	null	-0.947404	-0.320039	healthy	healthy	null	null	rule	null
2010-09-13	34	781.075	1,021.883188	0.750526	0.754906	0.947917	null	null	false	0.994198	0.960675	1.256855	0.025345	0.75	0.837772	1.120197	0	0.813772	0.742071	38.676667	-0.794361	null	null	null	0.001591	0.851593	null	null	-0.952769	-0.303695	healthy	healthy	null	null	rule	null
2010-09-14	34	574.6	1,026.278033	0.560253	0.751977	0.9375	null	null	true	0.74504	0.710261	1.268578	0.037731	1.25	0.862793	1.148029	2.5	0.830907	0.749006	36.477778	-0.661816	null	null	null	0.002163	12.118779	null	null	-0.957852	-0.287261	outage	outage	null	null	rule	actual <5% of expected for 10 consecutive intervals (~2.5h) with expected >20% of capacity
2010-09-15	34	-4.675	1,022.587676	-0.004773	0.751977	0.916667	null	null	true	-0.006347	-0.005808	1.270403	0.000369	0	0	1.157276	7.75	-0.002184	-0.001486	37.31	-0.205254	null	null	null	-0.00322	1.061281	null	null	-0.962652	-0.270742	outage	outage	null	null	rule	actual <5% of expected for 31 consecutive intervals (~7.8h) with expected >20% of capacity
2010-09-16	34	-2.725	953.414919	null	0.751977	0.71875	null	null	true	null	-0.003401	1.19009	0.00064	0	0	1.164672	4.5	-0.002816	-0.00147	36.978333	0.868299	null	null	null	-0.024313	0.776085	null	null	-0.967167	-0.254143	data_gap	data_gap	null	null	rule	only 72% of intervals judgeable — meter or sensor gap, performance unknown
2010-09-17	34	645.425	1,009.430197	0.694767	0.751977	0.864583	null	null	false	0.923921	0.809534	1.266092	0.019826	2	0.801453	1.135177	0.25	0.756847	0.706819	37.782778	0.479775	null	null	null	-0.024313	0.776085	null	null	-0.971395	-0.237468	healthy	healthy	null	null	rule	null
2010-09-18	34	765.0375	1,029.235248	0.72764	0.751252	0.895833	null	null	false	0.96857	0.964267	1.297266	0.026962	2.25	0.813559	1.154131	0	0.802467	0.707255	37.04	-0.766565	null	null	null	-0.022685	29.646269	null	null	-0.975336	-0.220724	healthy	healthy	null	null	rule	null
2010-09-19	34	750.775	1,015.447789	0.723178	0.750526	0.895833	null	null	false	0.963562	0.951007	1.286268	0.027651	2	0.803067	1.144238	0	0.800753	0.699819	39.065	-0.504552	null	null	null	-0.020304	null	null	null	-0.978989	-0.203914	healthy	healthy	null	null	rule	null
2010-09-20	34	757.075	965.174802	null	0.750076	0.791667	null	null	true	null	0.963843	1.228777	0.026527	1.75	0.801453	1.132587	0	0.809514	0.70868	37.293333	-0.637247	null	null	null	-0.017676	16.303496	null	null	-0.982352	-0.187043	data_gap	data_gap	null	null	rule	only 79% of intervals judgeable — meter or sensor gap, performance unknown
2010-09-21	34	755.95	986.131007	0.750331	0.750076	0.979167	null	null	false	1.000341	0.967358	1.261911	0.015736	2.25	0.822437	1.104756	0	0.807548	0.746677	37.298889	-0.429413	null	null	null	-0.017676	16.303496	null	null	-0.985424	-0.170118	healthy	healthy	null	null	rule	null
2010-09-22	34	791.45	1,003.601755	0.782756	0.750331	0.885417	null	null	false	1.043213	1.01807	1.290969	0.01379	1.5	0.881356	1.117331	0	0.834899	0.778356	30.728889	-0.519925	null	null	null	-0.013239	null	null	null	-0.988204	-0.153142	healthy	healthy	null	null	rule	null
2010-09-23	34	736.9	988.817452	0.73909	0.750429	0.958333	null	null	false	0.98489	0.95292	1.278686	0.025761	2	0.801453	1.110304	0	0.817447	0.718637	33.181111	-0.766104	null	null	null	-0.008113	null	null	null	-0.990692	-0.13612	healthy	healthy	null	null	rule	null
2010-09-24	34	726.775	975.54266	0.730516	0.750331	0.958333	null	null	false	0.973592	0.944876	1.268297	0.022603	1.75	0.788539	1.108454	0	0.825245	0.708754	35.633889	-0.690751	null	null	null	-0.003552	12.798342	null	null	-0.992887	-0.119059	healthy	healthy	null	null	rule	null
2010-09-25	34	724.65	977.521714	0.727554	0.749978	0.979167	null	null	false	0.9701	0.947247	1.277795	0.017881	2	0.795803	1.115944	0	0.808207	0.714671	36.876667	-0.659826	null	null	null	0.0012	12.023441	null	null	-0.994788	-0.101962	healthy	healthy	null	null	rule	null
2010-09-26	34	703.925	963.739812	0.717076	0.749626	0.979167	null	null	false	0.956579	0.92524	1.266741	0.023626	1.75	0.772397	1.102814	0	0.800364	0.702136	38.378889	-0.66508	null	null	null	0.00591	17.453505	null	null	-0.996395	-0.084835	healthy	healthy	null	null	rule	null
2010-09-27	34	682.975	956.465178	0.709017	0.749197	0.916667	null	null	false	0.946369	0.902731	1.26422	0.01894	2	0.751412	1.089037	0	0.797013	0.688685	39.599444	-0.913408	null	null	null	0.009962	18.102585	null	null	-0.997707	-0.067683	healthy	healthy	null	null	rule	null
2010-09-28	34	659.95	917.250533	0.715426	0.748529	0.979167	null	null	false	0.955776	0.877249	1.21927	0.026266	2	0.75222	1.085986	0	0.773741	0.690198	39.683889	-0.905845	null	null	null	0.013587	14.213637	null	null	-0.998724	-0.050511	healthy	healthy	null	null	rule	null
2010-09-29	34	650.9	899.118028	0.723932	0.743914	1	null	null	false	0.973139	0.870195	1.20204	0.027342	1.75	0.745763	1.056767	0	0.791218	0.7021	38.091667	-0.843771	null	null	null	0.017584	16.097601	null	null	-0.999445	-0.033324	healthy	healthy	null	null	rule	null
2010-09-30	34	404	536.398738	0.753171	0.739315	1	null	null	false	1.018742	0.543257	0.721293	0.14617	0.25	0.826473	1.131201	0	0.896234	0.749447	38.243889	-0.417144	null	null	null	0.022029	null	null	null	-0.99987	-0.016127	healthy	healthy	null	null	rule	null
2010-10-01	34	346.95	433.089391	0.780035	0.739315	0.989583	null	null	false	1.055079	0.469292	0.585806	0.150976	0.25	0.761098	1.007299	0	0.85152	0.788478	34.497222	-0.545379	null	null	null	0.021153	null	null	null	-0.999999	0.001075	healthy	healthy	null	null	rule	null
2010-10-02	34	517.6	693.604355	0.746247	0.743689	1	null	null	false	1.003439	0.704294	0.943781	0.053172	0.25	0.807103	1.082287	0	0.795887	0.75008	36.955556	-0.907763	null	null	null	0.001569	null	null	null	-0.999833	0.018277	healthy	healthy	null	null	rule	null
2010-10-03	34	432.275	536.573445	0.806646	0.747268	0.90625	null	null	false	1.079461	0.591742	0.734516	0.06031	0.25	1.078289	1.224404	0	0.875083	0.762862	28.956667	-0.477058	null	null	null	0.000633	null	null	null	-0.999371	0.035473	healthy	healthy	null	null	rule	null
2010-10-04	34	603.275	762.846004	0.73701	0.748529	0.90625	null	null	false	0.984611	0.830862	1.050631	0.071422	0.25	0.915254	1.131201	0	0.804875	0.726615	32.18	-0.045814	null	null	null	0.002279	null	null	null	-0.998613	0.052658	healthy	healthy	null	null	rule	null
2010-10-05	34	678.075	860.453078	0.788044	0.748529	1	null	null	false	1.05279	0.93964	1.19237	0.03815	1.75	0.806295	1.05011	0	0.841141	0.767141	24.704444	-0.633812	null	null	null	0.001441	5.227371	null	null	-0.997559	0.069828	healthy	healthy	null	null	rule	null
2010-10-06	34	606.15	787.041486	0.770163	0.748529	1	null	null	false	1.028902	0.845205	1.097437	0.035095	0.25	0.972559	1.064072	0	0.832369	0.774233	25.475	-0.616257	null	null	null	0.002896	null	null	null	-0.99621	0.086977	healthy	healthy	null	null	rule	null
2010-10-07	34	673.9	815.504574	0.820228	0.748529	0.854167	null	null	false	1.095787	0.94559	1.144284	0.061318	0.5	0.871671	1.064349	0	0.89791	0.787905	24.929444	-0.778112	null	null	null	0.004743	null	null	null	-0.994567	0.104101	healthy	healthy	null	null	rule	null
2010-10-08	34	682.625	887.123462	0.763676	0.748529	0.895833	null	null	false	1.020236	0.963921	1.252689	0.02571	1.75	0.784504	1.038367	0	0.830762	0.748367	27.601111	-0.586795	null	null	null	0.006636	null	null	null	-0.992629	0.121193	healthy	healthy	null	null	rule	null
2010-10-09	34	669.025	843.162872	0.754154	0.748529	0.958333	null	null	false	1.007515	0.950778	1.198252	0.019995	2	0.763519	1.029491	0	0.818788	0.741647	29.547778	-0.58021	null	null	null	0.006275	null	null	null	-0.990397	0.13825	healthy	healthy	null	null	rule	null
2010-10-10	34	676.325	863.457526	0.777282	0.748529	0.9375	null	null	false	1.038413	0.967375	1.235038	0.010753	1.25	0.803874	1.016361	0	0.817034	0.77624	29.93	-0.841966	null	null	null	0.005397	null	null	null	-0.987873	0.155266	healthy	healthy	null	null	rule	null
2010-10-11	34	638.85	861.604583	0.741465	0.748529	1	null	null	false	0.990563	0.91974	1.240435	0.015229	1.75	0.737692	0.999903	0	0.806851	0.727131	32.416111	-0.748225	null	null	null	0.004719	null	null	null	-0.985056	0.172236	healthy	healthy	null	null	rule	null
2010-10-12	34	641.775	852.854897	0.745203	0.747508	0.927083	null	null	false	0.996916	0.930031	1.235918	0.033052	2	0.742534	1.011923	0	0.805094	0.734659	31.229444	-0.528524	null	null	null	0.002486	null	null	null	-0.981947	0.189155	healthy	healthy	null	null	rule	null
2010-10-13	34	647.4	866.161303	0.739787	0.745725	0.96875	null	null	false	0.992038	0.944403	1.263523	0.020891	1.75	0.750605	1.027087	0	0.816104	0.723312	31.780556	-0.704649	null	null	null	0.000429	null	null	null	-0.978548	0.206018	healthy	healthy	null	null	rule	null
2010-10-14	34	613.5	784.687122	0.754919	0.743334	0.895833	null	null	false	1.015585	0.900927	1.152316	0.029015	1.75	0.732042	0.993799	0	0.810085	0.733941	32.343333	-0.582959	null	null	null	-0.001633	5.203005	null	null	-0.97486	0.22282	healthy	healthy	null	null	rule	null
2010-10-15	34	545.4625	731.450028	0.745933	0.745203	0.895833	null	null	false	1.000979	0.806427	1.081396	0.022763	2	0.716707	0.985478	0	0.796388	0.719819	33.299444	-0.758599	null	null	null	-0.002369	2.772535	null	null	-0.970882	0.239557	healthy	healthy	null	null	rule	null
2010-10-16	34	522.35	680.355226	0.768091	0.745568	0.895833	null	null	false	1.03021	0.777993	1.013328	0.036367	1.75	0.725585	0.988529	0	0.803323	0.728289	32.365556	-0.607555	null	null	null	-0.002483	221.738578	null	null	-0.966618	0.256222	healthy	healthy	null	null	rule	null
2010-10-17	34	399.85	481.921714	0.829699	0.745933	1	null	null	false	1.112297	0.600352	0.723578	0.074919	0.25	0.791768	0.847892	0	0.858705	0.859986	29.958333	-0.343618	null	null	null	-0.002928	1.806164	null	null	-0.962068	0.272811	healthy	healthy	null	null	rule	null
2010-10-18	34	612.15	728.408479	0.79557	0.746247	0.833333	null	null	false	1.066095	0.926592	1.102569	0.028946	1.25	0.762712	0.97318	0	0.860048	0.780705	25.750556	-0.934661	null	null	null	0.000282	null	null	null	-0.957232	0.28932	healthy	healthy	null	null	rule	null
2010-10-19	34	552.6875	635.90785	0.821837	0.750331	0.90625	null	null	false	1.095299	0.843442	0.970443	0.062511	0.25	0.932203	0.968188	0	0.828646	0.792785	28.484444	0.115617	null	null	null	0.00012	null	null	null	-0.952114	0.305743	healthy	healthy	null	null	rule	null
2010-10-20	34	10.45	379.769035	0.032622	0.753171	0.864583	null	null	true	0.043313	0.016079	0.584339	0.002747	0	0.297821	0.997036	2.5	-0.003299	-0.003894	23.978889	-0.796819	null	null	null	0.002133	null	null	null	-0.946714	0.322076	outage	outage	null	null	rule	actual <5% of expected for 10 consecutive intervals (~2.5h) with expected >20% of capacity
2010-10-21	34	-4.675	646.693897	-0.007657	0.753171	0.9375	null	null	true	-0.010166	-0.007253	1.003313	0.000863	0	0	0.970406	6.5	-0.004722	-0.001875	23.129444	0.727035	null	null	null	-0.019141	1.002042	null	null	-0.941034	0.338313	outage	outage	null	null	rule	actual <5% of expected for 26 consecutive intervals (~6.5h) with expected >20% of capacity
2010-10-22	34	-1.55	341.846054	null	0.753663	0.604167	null	null	true	null	-0.002425	0.534789	0	0	0	0.928798	0	null	null	22.516111	null	null	null	null	-0.036673	1.023182	null	null	-0.935075	0.35445	data_gap	data_gap	null	null	rule	only 60% of intervals judgeable — meter or sensor gap, performance unknown
2010-10-23	34	0	614.696667	null	0.753171	0.541667	null	null	true	null	0	0.969721	0	0	0	0.857323	0	null	null	23.545556	null	null	null	null	-0.036673	1.023182	null	null	-0.928839	0.370483	data_gap	data_gap	null	null	rule	only 54% of intervals judgeable — meter or sensor gap, performance unknown
2010-10-24	34	0	388.418654	null	0.753663	0.53125	null	null	true	null	0	0.617928	0	0	0	0.640958	0	null	null	23.511667	null	null	null	null	-0.036673	1.023182	null	null	-0.922329	0.386405	data_gap	data_gap	null	null	rule	only 53% of intervals judgeable — meter or sensor gap, performance unknown
2010-10-25	34	0	580.539959	null	0.754154	0.541667	null	null	true	null	0	0.931407	0	0	0	0.982612	0	null	null	22.336111	null	null	null	null	-0.036673	1.023182	null	null	-0.915546	0.402214	data_gap	data_gap	null	null	rule	only 54% of intervals judgeable — meter or sensor gap, performance unknown
2010-10-26	34	0	725.321003	null	0.754537	0.479167	null	null	true	null	0	1.173605	0	0	0	0.981872	0	null	null	21.273333	null	null	null	null	-0.036673	1.023182	null	null	-0.908492	0.417903	data_gap	data_gap	null	null	rule	only 48% of intervals judgeable — meter or sensor gap, performance unknown
2010-10-27	34	0	720.497985	null	0.754919	0.552083	null	null	true	null	0	1.175779	0	0	0	0.963194	0	null	null	17.965556	null	null	null	null	-0.036673	1.023182	null	null	-0.901169	0.433469	data_gap	data_gap	null	null	rule	only 55% of intervals judgeable — meter or sensor gap, performance unknown
2010-10-28	34	0	686.493019	null	0.759297	0.510417	null	null	true	null	0	1.129895	0	0	0	0.993522	0	null	null	21.273333	null	null	null	null	-0.036673	1.023182	null	null	-0.893579	0.448906	data_gap	data_gap	null	null	rule	only 51% of intervals judgeable — meter or sensor gap, performance unknown
2010-10-29	34	0	653.891252	null	0.763676	0.53125	null	null	true	null	0	1.085493	0	0	0	0.978913	0	null	null	24.310556	null	null	null	null	-0.036673	1.023182	null	null	-0.885725	0.46421	data_gap	data_gap	null	null	rule	only 53% of intervals judgeable — meter or sensor gap, performance unknown
2010-10-30	34	0	623.549412	null	0.765884	0.510417	null	null	true	null	0	1.044058	0	0	0	0.960143	0	null	null	21.070556	null	null	null	null	-0.036673	1.023182	null	null	-0.877609	0.479378	data_gap	data_gap	null	null	rule	only 51% of intervals judgeable — meter or sensor gap, performance unknown
2010-10-31	34	0	656.881271	null	0.768091	0.5625	null	null	true	null	0	1.109363	0	0	0	0.953116	0	null	null	22.701667	null	null	null	null	-0.036673	1.023182	null	null	-0.869233	0.494403	data_gap	data_gap	null	null	rule	only 56% of intervals judgeable — meter or sensor gap, performance unknown
2010-11-01	34	-0.2	630.540709	null	0.765884	0.604167	null	null	true	null	-0.000341	1.074084	0.000037	0	0	0.941927	0.75	null	-0.001724	26.105	null	null	null	null	-0.036673	1.023182	null	null	-0.8606	0.509282	data_gap	data_gap	null	null	rule	only 60% of intervals judgeable — meter or sensor gap, performance unknown
2010-11-02	34	-0.05	654.969686	null	0.768091	0.572917	null	null	true	null	-0.000086	1.125348	0	0	0	0.935409	0.25	null	-0.001726	27.556111	null	null	null	null	-0.036673	1.023182	null	null	-0.851712	0.52401	data_gap	data_gap	null	null	rule	only 57% of intervals judgeable — meter or sensor gap, performance unknown
2010-11-03	34	0	677.07876	null	0.765884	0.416667	null	null	true	null	0	1.173397	0	0	0	0.931386	0	null	null	29.305556	null	null	null	null	-0.036673	1.023182	null	null	-0.842572	0.538584	data_gap	data_gap	null	null	rule	only 42% of intervals judgeable — meter or sensor gap, performance unknown
2010-11-04	34	0	661.8304	null	0.768091	0.5	null	null	true	null	0	1.156884	0	0	0	0.919182	0	null	null	27.815	null	null	null	null	-0.036673	1.023182	null	null	-0.833183	0.552997	data_gap	data_gap	null	null	rule	only 50% of intervals judgeable — meter or sensor gap, performance unknown
2010-11-05	34	0	655.938991	null	0.768091	0.489583	null	null	true	null	0	1.15648	0	0	0	0.882936	0	null	null	27.241111	null	null	null	null	-0.036673	1.023182	null	null	-0.823547	0.567248	data_gap	data_gap	null	null	rule	only 49% of intervals judgeable — meter or sensor gap, performance unknown
2010-11-06	34	0	506.240316	null	0.768091	0.4375	null	null	true	null	0	0.900232	0	0	0	0.924822	0	null	null	25.885556	null	null	null	null	-0.036673	1.023182	null	null	-0.813668	0.58133	data_gap	data_gap	null	null	rule	only 44% of intervals judgeable — meter or sensor gap, performance unknown
2010-11-07	34	0	586.62895	null	0.768091	0.510417	null	null	true	null	0	1.05214	0	0	0	0.98557	0	null	null	24.108333	null	null	null	null	-0.036673	1.023182	null	null	-0.803548	0.59524	data_gap	data_gap	null	null	rule	only 51% of intervals judgeable — meter or sensor gap, performance unknown
2010-11-08	34	0	393.788669	null	0.768091	0.5625	null	null	true	null	0	0.712314	0	0	0	0.889223	0	null	null	18.376111	null	null	null	null	-0.036673	1.023182	null	null	-0.79319	0.608975	data_gap	data_gap	null	null	rule	only 56% of intervals judgeable — meter or sensor gap, performance unknown
2010-11-09	34	0	632.605517	null	0.768091	0.5625	null	null	true	null	0	1.154045	0	0	0	0.928798	0	null	null	15.923889	null	null	null	null	-0.036673	1.023182	null	null	-0.782597	0.622529	data_gap	data_gap	null	null	rule	only 56% of intervals judgeable — meter or sensor gap, performance unknown
2010-11-10	34	238.975	676.240116	null	0.768091	0.75	null	null	true	null	0.439647	1.244093	0.1758	0.75	0.7046	0.913125	0.25	null	0.78235	15.721111	0.728974	null	null	null	-0.036673	1.023182	null	null	-0.771773	0.635898	data_gap	data_gap	null	null	rule	only 75% of intervals judgeable — meter or sensor gap, performance unknown
2010-11-11	34	584	688.847148	0.806166	0.768091	0.96875	null	null	false	1.04957	1.08345	1.277965	0.020383	1.75	0.757062	0.915391	0	0.865387	0.805004	16.469444	-0.392793	null	null	null	-0.036673	1.023182	null	null	-0.76072	0.64908	healthy	healthy	null	null	rule	null
2010-11-12	34	549.825	696.560148	0.765835	0.768091	0.979167	null	null	false	0.997062	1.028568	1.303068	0.02231	1.5	0.719128	0.909288	0.25	0.835302	0.775602	16.480556	-0.420209	null	null	null	-0.029178	null	null	null	-0.749442	0.66207	healthy	healthy	null	null	rule	null
2010-11-13	34	568.7875	683.002987	0.790498	0.768091	0.947917	null	null	false	1.029172	1.072861	1.288296	0.018133	1.75	0.738499	0.907438	0	0.828354	0.800517	18.280556	-0.264427	null	null	null	-0.023336	68.943813	null	null	-0.737943	0.674863	healthy	healthy	null	null	rule	null
2010-11-14	34	533.775	649.117093	0.797969	0.768091	0.96875	null	null	false	1.038898	1.015092	1.23444	0.031175	0.5	0.71025	0.881641	0	0.813889	0.802499	19.911667	-0.089817	null	null	null	-0.016078	null	null	null	-0.726225	0.687457	healthy	healthy	null	null	rule	null
2010-11-15	34	543.8375	670.71781	0.788092	0.768091	0.958333	null	null	false	1.02604	1.042599	1.285843	0.0158	1.5	0.717514	0.880994	0	0.837398	0.802701	22.505	-0.302756	null	null	null	-0.009846	null	null	null	-0.714292	0.699848	healthy	healthy	null	null	rule	null
2010-11-16	34	522.375	641.495685	0.755539	0.768091	0.958333	null	null	false	0.983658	1.007682	1.237471	0.022122	2.25	0.673123	0.878775	0	0.808004	0.758632	24.462778	-0.28641	null	null	null	-0.004021	null	null	null	-0.702148	0.712031	healthy	healthy	null	null	rule	null
2010-11-17	34	523.525	669.772364	0.759512	0.768091	0.979167	null	null	false	0.98883	1.016068	1.299907	0.016662	1.5	0.68523	0.883121	0	0.80415	0.766432	19.265	-0.306165	null	null	null	0.000506	null	null	null	-0.689796	0.724004	healthy	healthy	null	null	rule	null
2010-11-18	34	523.425	680.598824	0.769065	0.768091	1	null	null	false	1.001268	1.021926	1.32879	0.017772	0.5	0.717514	0.904295	0	0.807576	0.774877	19.737778	-0.467131	null	null	null	0.005577	null	null	null	-0.67724	0.735762	healthy	healthy	null	null	rule	null
2010-11-19	34	323.275	374.742597	0.863593	0.768091	0.947917	null	null	false	1.124336	0.634829	0.735898	0.055563	0.5	0.726392	0.675077	0	0.9189	0.824621	20.997778	-0.531798	null	null	null	0.010604	null	null	null	-0.664483	0.747303	healthy	healthy	null	null	rule	null
2010-11-20	34	-4.675	206.29544	-0.020844	0.768091	0.90625	null	null	true	-0.027137	-0.009233	0.407408	0.001104	0	0	0.496807	1.5	-0.006184	-0.006533	17.268333	-0.123479	null	null	null	0.018961	null	null	null	-0.65153	0.758623	outage	outage	null	null	rule	actual <5% of expected for 6 consecutive intervals (~1.5h) with expected >20% of capacity
2010-11-21	34	-5	465.453191	-0.011138	0.768091	0.9375	null	null	true	-0.0145	-0.009929	0.92428	0.001228	0	0	0.887744	3.5	-0.004896	-0.002791	12.666667	0.4814	null	null	null	0.00366	1.083319	null	null	-0.638384	0.769718	outage	outage	null	null	rule	actual <5% of expected for 14 consecutive intervals (~3.5h) with expected >20% of capacity
2010-11-22	34	-5.6	634.948385	-0.008467	0.768091	0.9375	null	null	true	-0.011023	-0.01118	1.267604	0.000947	0	0	0.879052	6.75	-0.003763	-0.002784	13.325	0.54947	null	null	null	-0.008751	1.061781	null	null	-0.62505	0.780585	outage	outage	null	null	rule	actual <5% of expected for 27 consecutive intervals (~6.8h) with expected >20% of capacity
2010-11-23	34	238.2375	614.673733	0.449592	0.768091	0.895833	null	null	true	0.585336	0.478078	1.233483	0.124309	0.75	0.720743	0.863056	1.25	-0.005131	0.830028	13.628889	0.620638	null	null	null	-0.016673	21.524678	null	null	-0.61153	0.791221	outage	outage	null	null	rule	actual <5% of expected for 5 consecutive intervals (~1.2h) with expected >20% of capacity
2010-11-24	34	25	258.914124	0.001448	0.768091	0.9375	null	null	true	0.001886	0.050419	0.522169	0.003471	0	0.331719	0.773274	1.25	null	-0.004492	10.754444	-0.467861	null	null	null	-0.033877	null	null	null	-0.597829	0.801624	outage	outage	null	null	rule	actual <5% of expected for 5 consecutive intervals (~1.2h) with expected >20% of capacity
2010-11-25	34	550.25	635.821919	0.805137	0.768091	0.9375	null	null	false	1.048231	1.115068	1.288477	0.024902	1.75	0.73527	0.894863	0	0.86361	0.803198	7.767778	-0.119668	null	null	null	-0.067045	1.099065	null	null	-0.583951	0.811789	healthy	healthy	null	null	rule	null
2010-11-26	34	521.925	618.242335	0.784886	0.768091	0.958333	null	null	false	1.021866	1.06258	1.258671	0.021154	1.75	0.698951	0.871101	0	0.837918	0.792745	9.59	-0.056277	null	null	null	-0.050461	null	null	null	-0.569901	0.821713	healthy	healthy	null	null	rule	null
2010-11-27	34	396.75	429.179852	0.844634	0.768091	0.958333	null	null	false	1.099653	0.811303	0.877618	0.16429	0.25	0.754641	0.89098	0	0.821788	0.838666	11.333889	0.033406	null	null	null	-0.035781	157.608437	null	null	-0.555682	0.831395	healthy	healthy	null	null	rule	null
2010-11-28	34	514.025	482.21649	0.857499	0.768091	0.927083	null	null	false	1.116402	1.055539	0.99022	0.28307	0.25	0.837772	0.884415	0	0.864324	0.897126	11.530556	-0.120575	null	null	null	-0.018665	null	null	null	-0.541298	0.840831	healthy	healthy	null	null	rule	null
2010-11-29	34	537.625	647.911155	0.821949	0.768091	0.989583	null	null	false	1.070119	1.108411	1.335785	0.0414	1.5	0.744956	0.866385	0	0.872711	0.846635	8.678889	0.016378	null	null	null	-0.001173	null	null	null	-0.526755	0.850017	healthy	healthy	null	null	rule	null
2010-11-30	34	481.6	608.839602	0.791013	0.768091	1	null	null	false	1.029842	0.996652	1.259969	0.042942	0.25	0.691687	0.83356	0	0.830222	0.800059	9.595556	-0.011557	null	null	null	0.014717	null	null	null	-0.512055	0.858953	healthy	healthy	null	null	rule	null

End of preview. Expand in Data Studio

PVDAQ daily solar fault triage

One row per plant-day across 32 public PVDAQ photovoltaic systems: a shape summary of the day's power curve (24 engineered features), a fault label, and the label's provenance. 91,046 site-days, 1994–2025, from ~20 kW rooftops to a 38.7 MW farm, spanning desert, snow, coastal, and Southern-California (CAISO curtailment) climates.

This is a derived dataset. No raw PVDAQ time series are redistributed here: each row compresses one day of interval power, irradiance, and weather data into the checklist an expert would fill in from the daily curve — brightness vs the clear-sky ceiling, jaggedness, plateau length, temperature tracking, snow trail, trailing performance-index slope. Every feature is a ratio or a fraction of plant capacity, so a 2 kW roof and a 38 MW farm live on the same scale. NaN means "not measurable at this site" (no temperature sensor, short history), never zero.

Labels

final_label is the training target:

final_label	rows
healthy	54,531	unflagged day (negative class)
data_gap	19,753	too few judgeable intervals — performance unknowable; excluded from training
weather	4,967	broken-sky / rainy / deep-overcast underperformance
outage	3,875	full or partial production loss
unclassified	3,511	flagged but no rule reached a verdict — honest "don't know"; excluded from training
snow	1,923	snow-covered array
snow_shedding	964	morning snow-slide recovery pattern
cloud_intermittent	709	choppy broken-cloud day
clipping	394	inverter AC ceiling (found by shape census — chronic clipping hides inside rolling baselines)
curtailment	305	grid-commanded output limit (CAISO-region sites)
thermal	114	temperature-correlated derate

Excluding data_gap and unclassified leaves 67,782 rows across nine classes. final_label_2 (234 rows) carries the minor component of referee-split days — e.g. a weather day with a divergent-inverter loss riding under it.

Label provenance — the gold tier

Labels come from a rule classifier whose claims were then graded against per-inverter (sub-metered) data where it exists, by a fleet-relative referee:

provenance = "referee" (6,167 rows — the gold tier): the fleet actually graded the day. verdict is confirmed (4,241), attributed (1,578 — the referee overruled the rule label and final_label is the referee's attribution), or refuted (348).
provenance = "rule" (84,879 rows): weak labels resting on rule evidence alone. verdict is uncheckable (7,819) or empty (healthy / ungraded days).

evidence is a one-line human-readable justification for the day's label. In the source pipeline the gold tier is sample-weighted 5× during training.

Columns (37)

Identity and energy — date (local site date), site (PVDAQ system id, string), actual_kwh (measured AC energy), expected_kwh (modeled expected energy: on-site plane-of-array sensor where present, else Open-Meteo reanalysis irradiance, else pvlib clear-sky ceiling), pi (performance index = actual/expected), pi_baseline (trailing 30-day rolling baseline of pi, flagged days excluded), coverage (fraction of the day's intervals judgeable), rain_mm, snow_cm (daily precipitation/snowfall), flagged (day flagged as underperforming vs baseline).

Shape features — pi_rel (pi/pi_baseline), csr (clear-sky ratio of measured energy), model_csr (clear-sky ratio of the expected model — how dark the weather model thinks the day was), chop (intraday jaggedness, MAD-based), plateau_run_h (longest midday flat plateau in hours — the clipping/curtailment signature), peak_frac / expected_peak_frac (day peak power as fraction of AC capacity, actual and expected), dead_run_h (longest run of near-zero output while the model expects light, hours), morning_ratio / midday_ratio (median actual/expected ratio 07–10h and 11–14h), tmax_c (day maximum temperature), temp_corr (correlation of hourly performance ratio with temperature), snow_3d / snow_7d / snow_30d (trailing snowfall sums), pi_slope_14d (linear slope of pi over the trailing 14 days), pi_drop_share (largest single-day share of the trailing decline — step vs trend), unit_count / unit_dist (how the bright-hours deficit sits on the k-of-N inverter grid), doy_sin / doy_cos (day-of-year seasonality).

Labels — rule_label, final_label, final_label_2, verdict, provenance, evidence (see above).

Recommended splits

Faults are multi-day events. Consecutive same-label days at one site are one event — random row-level splits leak badly and overstate every metric. Either group consecutive (site, label) runs into events and split on events, or use leave-one-site-out, which is the headline protocol in the source pipeline (LOSO macro-F1 0.92 for a gradient-boosted model over these features).

Sites

32 PVDAQ system ids: 34, 35, 50, 51, 1199, 1200, 1202, 1203, 1239, 1276, 1277, 1278, 1283, 1318, 1325, 1367, 1368, 1403, 1418, 1419, 1420, 1423, 1433, 2107, 4902, 4903, 9069, 14597, 14601, 14645, 14695, 14697.

Highlights: 2107 "Farm Solar Array" (893 kW) and 9069 "Simon Solar Farm" (38.7 MW) are the two DOE Solar Data Bounty Prize plants; 50/51 are NREL's ~29-year crystalline-silicon test arrays (Golden, CO); 1278 is a deliberately overdriven clipping site (DC/AC 1.24); 14597/14601/14645 are a Southern-California CAISO trio (the curtailment source); 14695/14697 (Santa Ana, CA) were first used as a held-out validation pair. Authoritative per-system metadata (geometry, capacity, channels) lives in PVDAQ at csv/system_metadata/<id>_system_metadata.json.

A caution from onboarding: PVDAQ census fields and channel names are not always reliable (units flip mid-history on some systems, some azimuth fields are wrong). The features here were computed after per-site forensic validation of units, timestamps, and geometry.

How it was built

Per site: pvanalytics quality gate → canonical interval frame → expected power (POA trust ladder) → rolling-baseline flagging → rule classification → per-inverter referee grading → per-day feature export. prepare.py in this repo concatenates the 32 per-site exports into data/triage-days.parquet verbatim (no rows dropped beyond excluding the one non-PVDAQ site).

Sources and attribution

PVDAQ (NREL / OEDI). All underlying power, irradiance, and on-site sensor measurements come from the Photovoltaic Data Acquisition (PVDAQ) public datasets, hosted by the DOE Open Energy Data Initiative and licensed CC BY 4.0:

Deline, C., Perry, K., Deceglie, M., Muller, M., Sekulic, W., & Jordan, D. (2021). Photovoltaic Data Acquisition (PVDAQ) Public Datasets [Data set]. Open Energy Data Initiative (OEDI). NREL. https://doi.org/10.25984/1846021

OEDI submission: https://data.openei.org/submissions/4568
Data lake: s3://oedi-data-lake/pvdaq/
Systems 2107 and 9069 were contributed through the DOE American-Made Solar Data Bounty Prize (s3://oedi-data-lake/pvdaq/2023-solar-data-prize/); the same DOI covers them.

Open-Meteo / Copernicus. Weather-derived columns (rain_mm, snow_cm, the snow_*d trails) and, on sites without their own sensors, temperature and reanalysis-based expected power use the Open-Meteo Historical Weather API (CC BY 4.0), which serves ERA5. Weather data by Open-Meteo.com; contains modified Copernicus Climate Change Service information (ERA5).

Citation

If you use this dataset, please cite PVDAQ (above) and:

@misc{sn-triage-dataset,
  author = {Bernard, Logan},
  title  = {PVDAQ daily solar fault triage},
  year   = {2026},
  publisher = {Hugging Face},
  url    = {https://huggingface.co/datasets/hoodsy/sn-triage-dataset}
}

Downloads last month: 20

Models trained or fine-tuned on hoodsy/sn-triage-dataset

hoodsy/sn-triage

Tabular Classification • Updated 2 days ago