Hugging Face's logo

Datasets:
rajanie
/
DeepSolar-3M

Tags:
rooftop-solar
solar-energy
License:
Dataset card Data Studio Files
xet
Community
Dataset Preview
The full dataset viewer is not available (click to read why). Only showing a preview of the rows.
The dataset generation failed because of a cast error
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 27 new columns ({'Population Density (Per Sq. Mile)', 'cumulative num of residential PVs by 2009', 'Median household income ($)', 'cumulative num of residential PVs by 2016', 'cumulative num of residential PVs by 2014', 'cumulative num of residential PVs by 2011', 'cumulative num of residential PVs by 2013', 'cumulative num of residential PVs by 2022', 'Blkgrp details', 'Median Age', 'cumulative num of residential PVs by 2007', 'cumulative num of all PVs by 2022', 'cumulative num of residential PVs by 2008', 'state', 'Blkgrp_FIPS_code', 'Gini-diversity-index-2022', 'code', 'num of renter-occupied housing units', 'cumulative num of residential PVs by 2012', 'cumulative num of residential PVs by 2017', 'cumulative num of residential PVs by 2005', 'cumulative num of residential PVs by 2015', 'num of owner-occupied housing units', 'GHI (W/m^2)', 'cumulative num of residential PVs by 2010', 'cumulative num of residential PVs by 2006', "Population 25 Years and Over: Bachelor's Degree"}) and 9 missing columns ({'Average PV area (m^2)', 'pct residential systems (%)', 'Total PV area (m^2)', 'Total PV system count', 'pct commercial systems (%)', 'pct solar heat systems (%)', 'pct utility scale systems (%)', 'block_group_FIPS', 'Median PV area (m^2)'}).

This happened while the csv dataset builder was generating data using

hf://datasets/rajanie/DeepSolar-3M/dataset/blockgroup_level_historical_data.csv (at revision 43e9da5c5cceace5f867d222a51139a0be1c3cfc)

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 1831, in _prepare_split_single
                  writer.write_table(table)
                File "/usr/local/lib/python3.12/site-packages/datasets/arrow_writer.py", line 714, in write_table
                  pa_table = table_cast(pa_table, self._schema)
                             ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
                File "/usr/local/lib/python3.12/site-packages/datasets/table.py", line 2272, in table_cast
                  return cast_table_to_schema(table, schema)
                         ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
                File "/usr/local/lib/python3.12/site-packages/datasets/table.py", line 2218, in cast_table_to_schema
                  raise CastError(
              datasets.table.CastError: Couldn't cast
              cumulative num of residential PVs by 2005: double
              cumulative num of residential PVs by 2006: double
              cumulative num of residential PVs by 2007: double
              cumulative num of residential PVs by 2008: double
              cumulative num of residential PVs by 2009: double
              cumulative num of residential PVs by 2010: double
              cumulative num of residential PVs by 2011: double
              cumulative num of residential PVs by 2012: double
              cumulative num of residential PVs by 2013: double
              cumulative num of residential PVs by 2014: double
              cumulative num of residential PVs by 2015: double
              cumulative num of residential PVs by 2016: double
              cumulative num of residential PVs by 2017: double
              cumulative num of residential PVs by 2022: double
              cumulative num of all PVs by 2022: double
              GHI (W/m^2): double
              num of owner-occupied housing units: double
              num of renter-occupied housing units: double
              Median household income ($): double
              Population Density (Per Sq. Mile): double
              Blkgrp details: string
              Blkgrp_FIPS_code: int64
              Median Age: double
              state: string
              code: string
              Population 25 Years and Over: Bachelor's Degree: double
              Gini-diversity-index-2022: double
              -- schema metadata --
              pandas: '{"index_columns": [{"kind": "range", "name": null, "start": 0, "' + 4723
              to
              {'block_group_FIPS': Value('int64'), 'Total PV system count': Value('int64'), 'Total PV area (m^2)': Value('float64'), 'Median PV area (m^2)': Value('float64'), 'Average PV area (m^2)': Value('float64'), 'pct commercial systems (%)': Value('float64'), 'pct residential systems (%)': Value('float64'), 'pct solar heat systems (%)': Value('float64'), 'pct utility scale systems (%)': 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 1455, 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 1054, in convert_to_parquet
                  builder.download_and_prepare(
                File "/usr/local/lib/python3.12/site-packages/datasets/builder.py", line 894, in download_and_prepare
                  self._download_and_prepare(
                File "/usr/local/lib/python3.12/site-packages/datasets/builder.py", line 970, in _download_and_prepare
                  self._prepare_split(split_generator, **prepare_split_kwargs)
                File "/usr/local/lib/python3.12/site-packages/datasets/builder.py", line 1702, 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 1833, 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 27 new columns ({'Population Density (Per Sq. Mile)', 'cumulative num of residential PVs by 2009', 'Median household income ($)', 'cumulative num of residential PVs by 2016', 'cumulative num of residential PVs by 2014', 'cumulative num of residential PVs by 2011', 'cumulative num of residential PVs by 2013', 'cumulative num of residential PVs by 2022', 'Blkgrp details', 'Median Age', 'cumulative num of residential PVs by 2007', 'cumulative num of all PVs by 2022', 'cumulative num of residential PVs by 2008', 'state', 'Blkgrp_FIPS_code', 'Gini-diversity-index-2022', 'code', 'num of renter-occupied housing units', 'cumulative num of residential PVs by 2012', 'cumulative num of residential PVs by 2017', 'cumulative num of residential PVs by 2005', 'cumulative num of residential PVs by 2015', 'num of owner-occupied housing units', 'GHI (W/m^2)', 'cumulative num of residential PVs by 2010', 'cumulative num of residential PVs by 2006', "Population 25 Years and Over: Bachelor's Degree"}) and 9 missing columns ({'Average PV area (m^2)', 'pct residential systems (%)', 'Total PV area (m^2)', 'Total PV system count', 'pct commercial systems (%)', 'pct solar heat systems (%)', 'pct utility scale systems (%)', 'block_group_FIPS', 'Median PV area (m^2)'}).
              
              This happened while the csv dataset builder was generating data using
              
              hf://datasets/rajanie/DeepSolar-3M/dataset/blockgroup_level_historical_data.csv (at revision 43e9da5c5cceace5f867d222a51139a0be1c3cfc)
              
              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.

block_group_FIPS
int64
Total PV system count
int64
Total PV area (m^2)
float64
Median PV area (m^2)
float64
Average PV area (m^2)
float64
pct commercial systems (%)
float64
pct residential systems (%)
float64
pct solar heat systems (%)
float64
pct utility scale systems (%)
float64
10,010,201,002
1
60.609
60.609
60.609
0
100
0
0
10,010,203,001
1
22.006
22.006
22.006
0
100
0
0
10,010,204,001
1
17.281
17.281
17.281
0
100
0
0
10,010,204,002
1
14.981
14.981
14.981
0
100
0
0
10,010,205,002
1
11.314
11.314
11.314
0
100
0
0
10,010,205,003
1
96.415
96.415
96.415
0
100
0
0
10,010,208,021
1
29.59
29.59
29.59
0
100
0
0
10,010,209,001
1
50.539
50.539
50.539
0
100
0
0
10,010,210,002
1
32.325
32.325
32.325
0
0
100
0
10,010,211,001
1
6.962
6.962
6.962
0
100
0
0
10,010,211,002
1
28.222
28.222
28.222
100
0
0
0
10,030,101,002
1
32.2
32.2
32.2
0
100
0
0
10,030,102,002
1
46.249
46.249
46.249
0
100
0
0
10,030,103,001
2
15.914
7.957
7.957
0
100
0
0
10,030,103,002
6
159.448
22.006
26.575
0
66.667
33.333
0
10,030,104,001
1
64.277
64.277
64.277
0
0
100
0
10,030,104,002
2
67.82
33.91
33.91
50
0
50
0
10,030,104,003
1
34.563
34.563
34.563
0
0
100
0
10,030,105,001
2
61.293
30.646
30.646
0
50
50
0
10,030,105,003
1
9.449
9.449
9.449
0
100
0
0
10,030,106,003
1
21.508
21.508
21.508
0
0
100
0
10,030,107,011
1
46.933
46.933
46.933
0
0
100
0
10,030,107,012
1
25.922
25.922
25.922
0
100
0
0
10,030,107,014
1
3.295
3.295
3.295
0
100
0
0
10,030,107,031
10
232.365
20.234
23.237
0
50
50
0
10,030,107,032
3
78.077
17.468
26.026
0
100
0
0
10,030,107,051
3
460.379
34.314
153.46
33.333
66.667
0
0
10,030,107,052
1
32.76
32.76
32.76
0
100
0
0
10,030,107,053
1
9.822
9.822
9.822
0
100
0
0
10,030,108,001
2
57.128
28.564
28.564
0
100
0
0
10,030,108,002
3
44.074
15.789
14.691
0
100
0
0
10,030,108,003
2
47.928
23.964
23.964
0
100
0
0
10,030,109,031
1
44.882
44.882
44.882
0
100
0
0
10,030,109,032
1
36.801
36.801
36.801
0
100
0
0
10,030,109,033
2
31.703
15.852
15.852
0
0
100
0
10,030,109,042
2
51.968
25.984
25.984
0
100
0
0
10,030,109,051
4
97.782
22.472
24.446
0
75
25
0
10,030,109,052
1
23.933
23.933
23.933
0
100
0
0
10,030,110,001
1
108.785
108.785
108.785
100
0
0
0
10,030,110,002
3
138.81
45.13
46.27
0
66.667
33.333
0
10,030,111,012
1
27.6
27.6
27.6
0
100
0
0
10,030,111,013
3
88.769
23.373
29.59
0
66.667
33.333
0
10,030,111,021
2
55.698
27.849
27.849
0
100
0
0
10,030,112,011
2
67.944
33.972
33.972
0
100
0
0
10,030,113,001
1
20.638
20.638
20.638
0
100
0
0
10,030,113,002
3
70.431
23.373
23.477
33.333
66.667
0
0
10,030,113,003
1
48.239
48.239
48.239
0
100
0
0
10,030,113,004
2
35.433
17.716
17.716
0
100
0
0
10,030,114,011
2
41.214
20.607
20.607
0
100
0
0
10,030,114,012
2
112.391
56.195
56.195
0
50
50
0
10,030,114,014
2
59.801
29.9
29.9
0
100
0
0
10,030,114,015
3
107.418
35.619
35.806
0
100
0
0
10,030,114,031
1
20.452
20.452
20.452
0
0
100
0
10,030,114,032
6
289.68
51.626
48.28
0
66.667
33.333
0
10,030,114,033
4
157.832
37.64
39.458
0
50
50
0
10,030,114,051
8
270.658
22.845
33.832
0
75
25
0
10,030,114,052
23
773.495
27.787
33.63
0
91.304
8.696
0
10,030,114,062
3
98.591
35.744
32.864
0
66.667
33.333
0
10,030,114,063
1
14.049
14.049
14.049
0
100
0
0
10,030,114,071
8
182.076
17.468
22.759
0
62.5
37.5
0
10,030,114,072
7
616.719
68.752
88.103
14.286
85.714
0
0
10,030,114,073
3
61.106
15.914
20.369
0
33.333
66.667
0
10,030,115,011
7
283.526
40.219
40.504
14.286
71.429
14.286
0
10,030,115,012
1
14.67
14.67
14.67
0
0
100
0
10,030,115,013
1
101.637
101.637
101.637
100
0
0
0
10,030,115,021
10
473.682
41.307
47.368
10
70
20
0
10,030,115,022
1
52.901
52.901
52.901
0
100
0
0
10,030,115,023
1
7.646
7.646
7.646
0
100
0
0
10,030,115,024
4
171.259
47.399
42.815
0
100
0
0
10,030,116,012
4
114.318
26.015
28.579
0
75
25
0
10,030,116,013
3
99.15
31.641
33.05
0
33.333
66.667
0
10,030,116,014
3
96.539
25.611
32.18
0
33.333
66.667
0
10,030,116,021
4
183.816
32.014
45.954
0
100
0
0
10,030,116,022
3
149.626
49.917
49.875
0
100
0
0
10,030,116,023
3
61.604
23.373
20.535
0
100
0
0
10,030,116,024
3
132.967
58.06
44.322
0
33.333
66.667
0
10,059,501,003
2
24.306
12.153
12.153
0
100
0
0
10,059,502,003
1
113.448
113.448
113.448
100
0
0
0
10,059,505,001
1
19.768
19.768
19.768
0
100
0
0
10,059,507,001
2
103.812
51.906
51.906
0
50
50
0
10,059,508,001
1
109.407
109.407
109.407
0
0
100
0
10,059,509,002
1
21.198
21.198
21.198
0
100
0
0
10,070,100,021
1
28.16
28.16
28.16
0
100
0
0
10,070,100,023
2
27.414
13.707
13.707
50
50
0
0
10,070,100,032
1
15.106
15.106
15.106
0
100
0
0
10,070,100,042
1
24.244
24.244
24.244
0
100
0
0
10,090,501,012
1
146.394
146.394
146.394
0
100
0
0
10,090,501,014
1
103.253
103.253
103.253
0
0
100
0
10,090,501,015
2
36.987
18.494
18.494
0
100
0
0
10,090,501,021
1
20.016
20.016
20.016
0
0
100
0
10,090,501,022
3
124.823
27.476
41.608
66.667
0
33.333
0
10,090,501,023
3
230.749
83.361
76.916
66.667
33.333
0
0
10,090,502,001
2
59.677
29.838
29.838
0
50
50
0
10,090,503,001
2
33.817
16.908
16.908
0
50
50
0
10,090,503,002
2
13.862
6.931
6.931
50
50
0
0
10,090,503,003
1
36.925
36.925
36.925
100
0
0
0
10,090,503,004
3
71.177
25.922
23.726
0
100
0
0
10,090,504,003
1
62.66
62.66
62.66
100
0
0
0
10,090,505,001
3
158.702
57.376
52.901
33.333
66.667
0
0
10,090,505,004
2
59.552
29.776
29.776
50
50
0
0
End of preview.

DeepSolar-3M

🔨 Repo under active construction

📄 Paper: DeepSolar-3M: An AI-Enabled Solar PV Database Documenting 3 Million Systems Across the US
📍 Conference: ICLR 2025 - Tackling Climate Change with Machine Learning Workshop

Overview

DeepSolar-3M provides fast, high-resolution mapping of rooftop photovoltaic (PV) systems across the United States.
This repository contains county-level and blockgroup-level aggregated data from our AI pipeline.

Key features:

  • Scalable detection of PV installations from aerial imagery
  • Blockgroup-level and county-level aggregation of PV system statistics
  • Detailed breakdowns by system type (residential, commercial, utility-scale, solar heating)

📊 County-Level Dataset

Each entry corresponds to a U.S. county (identified by FIPS code) and includes:

  • Total PV system count
  • Total PV area (in square meters)
  • Median PV area (m²)
  • Average PV area (m²)

Breakdown by system type (% of systems):

  • Residential systems
  • Commercial systems
  • Utility-scale systems
  • Solar heating systems

🗺️ Block Group-Level Dataset

Each entry corresponds to a U.S. Census block group (identified by GEOID/Block Group FIPS) and includes all the features listed above.

📬 Citation

If you find this resource useful, please cite:

@inproceedings{prabha2025deepsolar3m,
  title={DeepSolar-3M: An AI-Enabled Solar PV Database Documenting 3 Million Systems Across the US},
  author={Prabha, Rajanie and Wang, Zhecheng and Zanocco, Chad and Flora, June and Rajagopal, Ram },
  booktitle={ICLR 2025 Workshop on Tackling Climate Change with Machine Learning},
  url={https://www.climatechange.ai/papers/iclr2025/55},
  year={2025}
}

Contact

Feel free to reach out in case you have any questions - [email protected]

Downloads last month
22