Datasets:

aiurban
/

cityshiftbench-scale122

Error code:   DatasetGenerationCastError
Exception:    DatasetGenerationCastError
Message:      An error occurred while generating the dataset

All the data files must have the same columns, but at some point there are 38 new columns ({'mean_delta_mae', 'median_delta_mae', 'holm_p', 'mean_delta_rmse', 'delta_rmse_ci95_lo', 'bootstrap_delta_mae_ci95_lo', 'mean_satca_r2', 'paired_t_p_delta_mae', 'wilcoxon_p_delta_rmse', 'median_delta_rmse', 'wilcoxon_p_delta_mae', 'bootstrap_ci95_hi', 'paired_city_count', 'paired_t_p_delta_r2', 'delta_mae_ci95_lo', 'bootstrap_ci95_lo', 'bootstrap_delta_r2_ci95_lo', 'holm_p_delta_rmse', 'bootstrap_delta_mae_ci95_hi', 'delta_r2_ci95_lo', 'delta_rmse_ci95_hi', 'negative_delta_r2_count', 'holm_p_delta_r2', 'holm_p_delta_mae', 'mean_source_r2', 'bootstrap_delta_rmse_ci95_lo', 'median_delta_r2', 'delta_mae_ci95_hi', 'mean_delta_r2', 'bootstrap_delta_rmse_ci95_hi', 'wilcoxon_p_delta_r2', 'bootstrap_delta_r2_ci95_hi', 'delta_r2_ci95_hi', 'zero_delta_r2_count', 'wilcoxon_p', 'comparison', 'paired_t_p_delta_rmse', 'positive_delta_r2_count'}) and 11 missing columns ({'positive_city_count', 'mean_r2', 'negative_city_count', 'cities_count', 'method_label', 'mean_rmse', 'mean_eval_count', 'method', 'negative_r2_ratio', 'positive_negative_cities', 'mean_mae'}).

This happened while the csv dataset builder was generating data using

hf://datasets/aiurban/cityshiftbench-scale122/results/satca_scale122_local_sharded_k010_paired_significance.csv (at revision dde267fd4eebdeb6905fb9b68daffbd3607ccb3d), [/tmp/hf-datasets-cache/medium/datasets/40379862974343-config-parquet-and-info-aiurban-cityshiftbench-sc-23c4707a/hub/datasets--aiurban--cityshiftbench-scale122/snapshots/dde267fd4eebdeb6905fb9b68daffbd3607ccb3d/results/satca_scale122_local_sharded_k010_paired_significance.csv (origin=hf://datasets/aiurban/cityshiftbench-scale122@dde267fd4eebdeb6905fb9b68daffbd3607ccb3d/results/satca_scale122_local_sharded_k010_paired_significance.csv)]

Please either edit the data files to have matching columns, or separate them into different configurations (see docs at https://hf.co/docs/hub/datasets-manual-configuration#multiple-configurations)
Traceback:    Traceback (most recent call last):
                File "/usr/local/lib/python3.12/site-packages/datasets/builder.py", line 1800, in _prepare_split_single
                  writer.write_table(table)
                File "/usr/local/lib/python3.12/site-packages/datasets/arrow_writer.py", line 765, in write_table
                  self._write_table(pa_table, writer_batch_size=writer_batch_size)
                File "/usr/local/lib/python3.12/site-packages/datasets/arrow_writer.py", line 773, in _write_table
                  pa_table = table_cast(pa_table, self._schema)
                             ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
                File "/usr/local/lib/python3.12/site-packages/datasets/table.py", line 2321, in table_cast
                  return cast_table_to_schema(table, schema)
                         ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
                File "/usr/local/lib/python3.12/site-packages/datasets/table.py", line 2249, in cast_table_to_schema
                  raise CastError(
              datasets.table.CastError: Couldn't cast
              slice_id: string
              target: string
              shot: int64
              comparison: string
              paired_city_count: int64
              mean_source_r2: double
              mean_satca_r2: double
              mean_delta_r2: double
              median_delta_r2: double
              delta_r2_ci95_lo: double
              delta_r2_ci95_hi: double
              bootstrap_delta_r2_ci95_lo: double
              bootstrap_delta_r2_ci95_hi: double
              bootstrap_ci95_lo: double
              bootstrap_ci95_hi: double
              positive_delta_r2_count: int64
              negative_delta_r2_count: int64
              zero_delta_r2_count: int64
              paired_t_p_delta_r2: double
              wilcoxon_p_delta_r2: double
              wilcoxon_p: double
              mean_delta_rmse: double
              median_delta_rmse: double
              delta_rmse_ci95_lo: double
              delta_rmse_ci95_hi: double
              bootstrap_delta_rmse_ci95_lo: double
              bootstrap_delta_rmse_ci95_hi: double
              paired_t_p_delta_rmse: double
              wilcoxon_p_delta_rmse: double
              mean_delta_mae: double
              median_delta_mae: double
              delta_mae_ci95_lo: double
              delta_mae_ci95_hi: double
              bootstrap_delta_mae_ci95_lo: double
              bootstrap_delta_mae_ci95_hi: double
              paired_t_p_delta_mae: double
              wilcoxon_p_delta_mae: double
              holm_p_delta_r2: double
              holm_p_delta_rmse: double
              holm_p_delta_mae: double
              holm_p: double
              -- schema metadata --
              pandas: '{"index_columns": [{"kind": "range", "name": null, "start": 0, "' + 5795
              to
              {'slice_id': Value('string'), 'target': Value('string'), 'shot': Value('int64'), 'method': Value('string'), 'method_label': Value('string'), 'cities_count': Value('int64'), 'positive_city_count': Value('int64'), 'negative_city_count': Value('int64'), 'positive_negative_cities': Value('string'), 'negative_r2_ratio': Value('float64'), 'mean_r2': Value('float64'), 'mean_rmse': Value('float64'), 'mean_mae': Value('float64'), 'mean_eval_count': Value('float64')}
              because column names don't match
              
              During handling of the above exception, another exception occurred:
              
              Traceback (most recent call last):
                File "/src/services/worker/src/worker/job_runners/config/parquet_and_info.py", line 1347, in compute_config_parquet_and_info_response
                  parquet_operations = convert_to_parquet(builder)
                                       ^^^^^^^^^^^^^^^^^^^^^^^^^^^
                File "/src/services/worker/src/worker/job_runners/config/parquet_and_info.py", line 980, in convert_to_parquet
                  builder.download_and_prepare(
                File "/usr/local/lib/python3.12/site-packages/datasets/builder.py", line 882, in download_and_prepare
                  self._download_and_prepare(
                File "/usr/local/lib/python3.12/site-packages/datasets/builder.py", line 943, in _download_and_prepare
                  self._prepare_split(split_generator, **prepare_split_kwargs)
                File "/usr/local/lib/python3.12/site-packages/datasets/builder.py", line 1646, in _prepare_split
                  for job_id, done, content in self._prepare_split_single(
                                               ^^^^^^^^^^^^^^^^^^^^^^^^^^^
                File "/usr/local/lib/python3.12/site-packages/datasets/builder.py", line 1802, in _prepare_split_single
                  raise DatasetGenerationCastError.from_cast_error(
              datasets.exceptions.DatasetGenerationCastError: An error occurred while generating the dataset
              
              All the data files must have the same columns, but at some point there are 38 new columns ({'mean_delta_mae', 'median_delta_mae', 'holm_p', 'mean_delta_rmse', 'delta_rmse_ci95_lo', 'bootstrap_delta_mae_ci95_lo', 'mean_satca_r2', 'paired_t_p_delta_mae', 'wilcoxon_p_delta_rmse', 'median_delta_rmse', 'wilcoxon_p_delta_mae', 'bootstrap_ci95_hi', 'paired_city_count', 'paired_t_p_delta_r2', 'delta_mae_ci95_lo', 'bootstrap_ci95_lo', 'bootstrap_delta_r2_ci95_lo', 'holm_p_delta_rmse', 'bootstrap_delta_mae_ci95_hi', 'delta_r2_ci95_lo', 'delta_rmse_ci95_hi', 'negative_delta_r2_count', 'holm_p_delta_r2', 'holm_p_delta_mae', 'mean_source_r2', 'bootstrap_delta_rmse_ci95_lo', 'median_delta_r2', 'delta_mae_ci95_hi', 'mean_delta_r2', 'bootstrap_delta_rmse_ci95_hi', 'wilcoxon_p_delta_r2', 'bootstrap_delta_r2_ci95_hi', 'delta_r2_ci95_hi', 'zero_delta_r2_count', 'wilcoxon_p', 'comparison', 'paired_t_p_delta_rmse', 'positive_delta_r2_count'}) and 11 missing columns ({'positive_city_count', 'mean_r2', 'negative_city_count', 'cities_count', 'method_label', 'mean_rmse', 'mean_eval_count', 'method', 'negative_r2_ratio', 'positive_negative_cities', 'mean_mae'}).
              
              This happened while the csv dataset builder was generating data using
              
              hf://datasets/aiurban/cityshiftbench-scale122/results/satca_scale122_local_sharded_k010_paired_significance.csv (at revision dde267fd4eebdeb6905fb9b68daffbd3607ccb3d), [/tmp/hf-datasets-cache/medium/datasets/40379862974343-config-parquet-and-info-aiurban-cityshiftbench-sc-23c4707a/hub/datasets--aiurban--cityshiftbench-scale122/snapshots/dde267fd4eebdeb6905fb9b68daffbd3607ccb3d/results/satca_scale122_local_sharded_k010_paired_significance.csv (origin=hf://datasets/aiurban/cityshiftbench-scale122@dde267fd4eebdeb6905fb9b68daffbd3607ccb3d/results/satca_scale122_local_sharded_k010_paired_significance.csv)]
              
              Please either edit the data files to have matching columns, or separate them into different configurations (see docs at https://hf.co/docs/hub/datasets-manual-configuration#multiple-configurations)

Need help to make the dataset viewer work? Make sure to review how to configure the dataset viewer, and open a discussion for direct support.

slice_id string	target string	shot int64	method string	method_label string	cities_count int64	positive_city_count int64	negative_city_count int64	positive_negative_cities string	negative_r2_ratio float64	mean_r2 float64	mean_rmse float64	mean_mae float64	mean_eval_count float64
scale122	intersection_nodes	0	coral_ridge	CORAL ridge	118	36	82	36/82	0.694915	-22.195125	20.413635	14.974253	70.838983
scale122	intersection_nodes	0	iw_ridge	Importance-weighted ridge	118	76	42	76/42	0.355932	-0.629996	11.330094	7.893193	70.838983
scale122	intersection_nodes	0	pooled_ridge	Pooled ridge	118	55	63	55/63	0.533898	-1.552785	12.71692	9.059955	70.838983
scale122	intersection_nodes	0	source_only_ridge	Source-only ridge	118	55	63	55/63	0.533898	-1.552785	12.71692	9.059955	70.838983
scale122	intersection_nodes	0	tabular_knn	Tabular kNN	118	92	26	92/26	0.220339	0.192524	10.046254	6.560225	70.838983
scale122	intersection_nodes	1	coral_ridge	CORAL ridge	118	35	83	35/83	0.70339	-22.228541	20.435937	14.995132	69.838983
scale122	intersection_nodes	1	iw_ridge	Importance-weighted ridge	118	76	42	76/42	0.355932	-0.628523	11.332763	7.896797	69.838983
scale122	intersection_nodes	1	pooled_ridge	Pooled ridge	118	55	63	55/63	0.533898	-1.55711	12.728478	9.072472	69.838983
scale122	intersection_nodes	1	source_only_ridge	Source-only ridge	118	55	63	55/63	0.533898	-1.558233	12.730687	9.074434	69.838983
scale122	intersection_nodes	1	tabular_knn	Tabular kNN	118	92	26	92/26	0.220339	0.188718	10.058207	6.567923	69.838983
scale122	intersection_nodes	1	target_only_ridge	Target-only ridge	118	0	118	0/118	1	-1.18351	17.63306	13.406471	69.838983
scale122	intersection_nodes	5	coral_ridge	CORAL ridge	118	35	83	35/83	0.70339	-24.352768	20.33104	14.945215	65.838983
scale122	intersection_nodes	5	iw_ridge	Importance-weighted ridge	118	75	43	75/43	0.364407	-0.716192	11.232678	7.843649	65.838983
scale122	intersection_nodes	5	pooled_ridge	Pooled ridge	118	53	65	53/65	0.550847	-1.73331	12.662083	9.045706	65.838983
scale122	intersection_nodes	5	source_only_ridge	Source-only ridge	118	53	65	53/65	0.550847	-1.739187	12.673903	9.055811	65.838983
scale122	intersection_nodes	5	tabular_knn	Tabular kNN	118	91	27	91/27	0.228814	0.171769	9.921179	6.510563	65.838983
scale122	intersection_nodes	5	target_only_ridge	Target-only ridge	118	73	45	73/45	0.381356	-0.067715	11.364416	7.756512	65.838983
scale122	intersection_nodes	10	coral_ridge	CORAL ridge	118	31	87	31/87	0.737288	-27.32437	20.254596	14.977625	60.838983
scale122	intersection_nodes	10	iw_ridge	Importance-weighted ridge	118	73	45	73/45	0.381356	-0.887737	11.099045	7.808659	60.838983
scale122	intersection_nodes	10	pooled_ridge	Pooled ridge	118	51	67	51/67	0.567797	-2.101072	12.583911	9.041922	60.838983
scale122	intersection_nodes	10	source_only_ridge	Source-only ridge	118	51	67	51/67	0.567797	-2.114589	12.60689	9.061938	60.838983
scale122	intersection_nodes	10	tabular_knn	Tabular kNN	118	88	30	88/30	0.254237	0.100772	9.780755	6.481869	60.838983
scale122	intersection_nodes	10	target_only_ridge	Target-only ridge	118	84	34	84/34	0.288136	-0.12682	10.429543	7.171289	60.838983
scale122	road_segments	0	coral_ridge	CORAL ridge	118	38	80	38/80	0.677966	-4.508795	374.879263	293.577535	70.838983
scale122	road_segments	0	iw_ridge	Importance-weighted ridge	118	82	36	82/36	0.305085	0.095395	223.274762	164.023991	70.838983
scale122	road_segments	0	pooled_ridge	Pooled ridge	118	77	41	77/41	0.347458	-0.125002	236.614327	176.057845	70.838983
scale122	road_segments	0	source_only_ridge	Source-only ridge	118	77	41	77/41	0.347458	-0.125002	236.614327	176.057845	70.838983
scale122	road_segments	0	tabular_knn	Tabular kNN	118	78	40	78/40	0.338983	0.004763	241.37878	168.938978	70.838983
scale122	road_segments	1	coral_ridge	CORAL ridge	118	37	81	37/81	0.686441	-4.5515	375.060485	293.82258	69.838983
scale122	road_segments	1	iw_ridge	Importance-weighted ridge	118	82	36	82/36	0.305085	0.095781	223.343902	164.15764	69.838983
scale122	road_segments	1	pooled_ridge	Pooled ridge	118	77	41	77/41	0.347458	-0.126705	236.837772	176.325933	69.838983
scale122	road_segments	1	source_only_ridge	Source-only ridge	118	77	41	77/41	0.347458	-0.12718	236.875044	176.358154	69.838983
scale122	road_segments	1	tabular_knn	Tabular kNN	118	77	41	77/41	0.347458	0.006524	241.600491	169.148384	69.838983
scale122	road_segments	1	target_only_ridge	Target-only ridge	118	0	118	0/118	1	-1.202709	384.188132	304.885791	69.838983
scale122	road_segments	5	coral_ridge	CORAL ridge	118	36	82	36/82	0.694915	-4.569638	373.960822	293.59085	65.838983
scale122	road_segments	5	iw_ridge	Importance-weighted ridge	118	82	36	82/36	0.305085	0.083831	222.639846	163.983936	65.838983
scale122	road_segments	5	pooled_ridge	Pooled ridge	118	73	45	73/45	0.381356	-0.151848	236.768763	176.71913	65.838983
scale122	road_segments	5	source_only_ridge	Source-only ridge	118	73	45	73/45	0.381356	-0.154315	236.968762	176.894551	65.838983
scale122	road_segments	5	tabular_knn	Tabular kNN	118	76	42	76/42	0.355932	-0.001338	240.369463	168.746212	65.838983
scale122	road_segments	5	target_only_ridge	Target-only ridge	118	63	55	63/55	0.466102	-0.093	256.972087	186.022573	65.838983
scale122	road_segments	10	coral_ridge	CORAL ridge	118	33	85	33/85	0.720339	-4.74778	373.012626	294.03593	60.838983
scale122	road_segments	10	iw_ridge	Importance-weighted ridge	118	82	36	82/36	0.305085	0.054226	221.128469	163.620703	60.838983
scale122	road_segments	10	pooled_ridge	Pooled ridge	118	73	45	73/45	0.381356	-0.201306	236.145156	176.962276	60.838983
scale122	road_segments	10	source_only_ridge	Source-only ridge	118	73	45	73/45	0.381356	-0.206587	236.551175	177.325889	60.838983
scale122	road_segments	10	tabular_knn	Tabular kNN	118	79	39	79/39	0.330508	-0.015554	238.245888	167.8319	60.838983
scale122	road_segments	10	target_only_ridge	Target-only ridge	118	86	32	86/32	0.271186	0.137182	228.798596	165.353252	60.838983

CityShiftBench Scale-122

CityShiftBench is an anonymous NeurIPS 2026 Evaluations and Datasets review artifact for low-shot cross-city urban regression under strict city isolation. The active Scale-122 surface contains 118 OSM-integrity-passing cities and 8,359 tile records. The paper-core targets are OSM-derived Road (target_road_segments) and Connectivity (target_intersection_nodes).

data/cityshiftbench_scale122_tile_targets.csv: tile-level target and target-safe descriptor table used by the benchmark.
data/tile_manifest_122.csv: tile registry and geometry descriptors.
data/splits/target_shot_registry_scale122.csv: fixed support/evaluation row assignments for shots 0,1,5,10,20 and seeds 7,19,42,61,97,123,211,307.
results/: released benchmark summaries, city-wise scores, paired significance files, control summaries, and diagnostics.
scripts/: executable experiment entry points.
docs/: data card, benchmark card, compute resources, license notes, artifact release notes, and Croissant metadata.
croissant_cityshiftbench_scale122.json: Croissant metadata with Responsible-AI fields for OpenReview upload.

Intended Use

Use this artifact to evaluate low-shot cross-city transfer while preserving the registered city universe, target definitions, shot indices, city-wise metrics, paired inference, and validity controls.

Not Intended For

The artifact is not an operational urban planning system and should not be used to rank cities or allocate resources without local validation.

Licenses

Code is released under MIT. Author-created benchmark metadata and generated result summaries are released under CC BY 4.0 unless otherwise stated. OpenStreetMap-derived target fields remain subject to ODbL 1.0 and require attribution to OpenStreetMap contributors.

Downloads last month: 90

CityShiftBench Scale-122

Contents

Intended Use

Not Intended For

Licenses