Dataset Preview
Duplicate
The full dataset viewer is not available (click to read why). Only showing a preview of the rows.
The dataset generation failed
Error code:   DatasetGenerationError
Exception:    ArrowInvalid
Message:      Float value 0.237480 was truncated converting to int64
Traceback:    Traceback (most recent call last):
                File "/usr/local/lib/python3.14/site-packages/datasets/builder.py", line 1837, in _prepare_split_single
                  writer.write_table(table)
                  ~~~~~~~~~~~~~~~~~~^^^^^^^
                File "/usr/local/lib/python3.14/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/python3.14/site-packages/datasets/arrow_writer.py", line 773, in _write_table
                  pa_table = table_cast(pa_table, self._schema)
                File "/usr/local/lib/python3.14/site-packages/datasets/table.py", line 2369, in table_cast
                  return cast_table_to_schema(table, schema)
                File "/usr/local/lib/python3.14/site-packages/datasets/table.py", line 2303, in cast_table_to_schema
                  cast_array_to_feature(
                  ~~~~~~~~~~~~~~~~~~~~~^
                      table[name] if name in table_column_names else pa.array([None] * len(table), type=schema.field(name).type),
                      ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
                      feature,
                      ^^^^^^^^
                  )
                  ^
                File "/usr/local/lib/python3.14/site-packages/datasets/table.py", line 1852, in wrapper
                  return pa.chunked_array([func(chunk, *args, **kwargs) for chunk in array.chunks])
                                           ~~~~^^^^^^^^^^^^^^^^^^^^^^^^
                File "/usr/local/lib/python3.14/site-packages/datasets/table.py", line 2143, in cast_array_to_feature
                  return array_cast(
                      array,
                  ...<2 lines>...
                      allow_decimal_to_str=allow_decimal_to_str,
                  )
                File "/usr/local/lib/python3.14/site-packages/datasets/table.py", line 1854, in wrapper
                  return func(array, *args, **kwargs)
                File "/usr/local/lib/python3.14/site-packages/datasets/table.py", line 2006, in array_cast
                  return array.cast(pa_type)
                         ~~~~~~~~~~^^^^^^^^^
                File "pyarrow/array.pxi", line 1147, in pyarrow.lib.Array.cast
                File "/usr/local/lib/python3.14/site-packages/pyarrow/compute.py", line 412, in cast
                  return call_function("cast", [arr], options, memory_pool)
                File "pyarrow/_compute.pyx", line 604, in pyarrow._compute.call_function
                File "pyarrow/_compute.pyx", line 399, in pyarrow._compute.Function.call
                  result = GetResultValue(
                File "pyarrow/error.pxi", line 155, in pyarrow.lib.pyarrow_internal_check_status
                File "pyarrow/error.pxi", line 92, in pyarrow.lib.check_status
                  raise convert_status(status)
              pyarrow.lib.ArrowInvalid: Float value 0.237480 was truncated converting to int64
              
              The above exception was the direct cause of the following exception:
              
              Traceback (most recent call last):
                File "/src/services/worker/src/worker/job_runners/config/parquet_and_info.py", line 1369, in compute_config_parquet_and_info_response
                  parquet_operations, partial, estimated_dataset_info = stream_convert_to_parquet(
                                                                        ~~~~~~~~~~~~~~~~~~~~~~~~~^
                      builder, max_dataset_size_bytes=max_dataset_size_bytes
                      ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
                  )
                  ^
                File "/src/services/worker/src/worker/job_runners/config/parquet_and_info.py", line 948, in stream_convert_to_parquet
                  builder._prepare_split(split_generator=splits_generators[split], file_format="parquet")
                  ~~~~~~~~~~~~~~~~~~~~~~^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
                File "/usr/local/lib/python3.14/site-packages/datasets/builder.py", line 1683, in _prepare_split
                  for job_id, done, content in self._prepare_split_single(
                                               ~~~~~~~~~~~~~~~~~~~~~~~~~~^
                      gen_kwargs=gen_kwargs, job_id=job_id, **_prepare_split_args
                      ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
                  ):
                  ^
                File "/usr/local/lib/python3.14/site-packages/datasets/builder.py", line 1869, in _prepare_split_single
                  raise DatasetGenerationError("An error occurred while generating the dataset") from e
              datasets.exceptions.DatasetGenerationError: An error occurred while generating the dataset

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.

f0
int64
f1
int64
f2
int64
f3
int64
f4
int64
f5
int64
f6
int64
f7
int64
f8
int64
f9
int64
label
int64
2,804
139
9
268
65
3,180
234
238
135
6,121
0
2,785
155
18
242
118
3,090
238
238
122
6,211
0
2,579
132
6
300
-15
67
230
237
140
6,031
0
2,886
151
11
371
26
5,253
234
240
136
4,051
0
2,742
134
22
150
69
3,215
248
224
92
6,091
0
2,880
209
17
216
30
4,986
206
253
179
4,323
0
2,962
148
16
323
23
5,916
240
236
120
3,395
0
2,811
135
1
212
30
3,670
220
238
154
5,643
0
2,900
45
19
242
20
5,199
221
195
100
4,115
0
2,570
346
2
0
0
331
215
235
158
5,745
0
2,678
128
5
95
23
1,660
229
236
141
6,546
0
2,952
107
11
42
7
5,845
239
226
116
3,509
0
2,705
90
8
134
22
2,023
232
228
129
6,615
0
2,740
54
6
218
42
2,287
224
227
138
6,686
0
2,640
80
8
180
-2
1,092
231
225
127
5,866
0
2,843
166
12
242
53
4,434
230
244
144
4,956
0
3,008
45
14
277
10
6,371
223
208
116
3,036
0
2,893
114
16
108
30
5,066
245
223
102
4,340
0
2,850
6
9
0
0
4,858
210
223
151
4,548
0
2,628
30
10
240
19
960
217
218
136
5,645
0
2,864
118
18
201
74
4,567
248
221
93
4,849
0
2,827
160
28
134
65
3,948
235
233
108
5,474
0
2,529
326
5
30
14
1,062
207
234
166
5,047
0
2,840
153
26
134
42
4,613
241
231
102
4,833
0
2,746
143
16
67
22
2,440
241
235
119
6,597
0
2,537
42
7
210
17
1,132
222
224
137
4,919
0
2,818
332
26
30
17
4,526
151
197
181
4,978
0
2,801
18
7
560
58
3,084
215
226
148
6,457
0
2,791
63
10
418
48
2,942
229
221
124
6,606
0
2,745
306
11
67
24
2,416
190
234
184
6,428
0
2,514
102
6
272
-5
1,082
230
233
137
4,811
0
2,788
13
16
30
8
4,126
203
206
137
5,396
0
2,562
354
12
67
9
1,057
200
218
156
5,031
0
3,073
173
12
108
-3
6,836
227
246
149
2,735
0
2,978
71
10
426
85
5,742
231
221
121
3,792
0
3,067
164
11
85
7
6,811
230
243
144
2,774
0
2,804
72
5
543
61
3,115
225
231
141
6,471
0
2,562
59
3
0
0
1,116
221
233
148
5,091
0
2,567
34
9
190
16
1,136
219
221
138
4,924
0
2,998
45
8
351
16
5,842
223
222
134
3,721
0
2,684
84
12
150
2
1,677
237
220
113
5,693
0
2,676
42
11
150
-5
1,634
222
216
128
5,655
0
2,574
170
5
170
23
1,180
224
242
153
4,910
0
2,511
211
6
162
30
1,071
216
245
166
4,660
0
3,067
32
4
30
-2
6,679
219
230
147
2,947
0
2,909
57
16
134
23
5,502
229
203
102
4,084
0
2,567
333
1
0
0
1,266
216
237
158
5,079
0
2,752
332
6
342
24
2,372
207
233
165
6,131
0
2,751
88
5
400
30
2,322
228
231
137
6,088
0
2,569
102
7
228
18
1,266
232
231
132
4,844
0
2,982
53
14
240
63
5,756
227
209
112
3,880
0
2,586
76
4
190
30
1,290
225
232
143
4,875
0
2,723
99
6
301
34
2,109
230
232
136
5,793
0
3,070
0
11
30
-6
6,890
204
220
153
2,858
0
3,057
124
12
150
53
6,508
240
231
118
3,211
0
2,818
119
19
30
10
5,213
248
220
92
4,497
0
2,847
320
33
85
39
4,983
120
190
199
4,727
0
2,606
112
11
30
4
1,584
239
228
119
5,106
0
2,893
83
6
551
101
4,011
229
230
135
5,755
0
2,716
1
10
234
27
2,100
206
222
153
5,581
0
2,644
125
20
67
25
1,719
249
221
90
5,168
0
2,613
37
5
216
34
1,519
220
229
146
4,818
0
2,929
356
12
0
0
5,757
201
219
155
4,017
0
2,767
307
12
30
8
2,796
187
233
186
6,192
0
2,939
56
18
95
20
5,563
229
200
98
4,224
0
2,537
7
12
0
0
1,583
205
216
148
4,509
0
3,119
119
7
162
10
6,660
233
234
133
2,556
0
2,775
344
13
42
7
2,854
194
219
164
6,177
0
2,490
79
9
240
13
878
232
224
124
4,184
0
2,831
284
24
60
23
5,208
145
232
222
4,611
0
2,837
112
8
272
16
3,649
235
231
128
6,221
0
2,890
299
23
180
79
5,104
150
224
212
4,734
0
2,824
135
5
218
6
3,583
227
238
144
6,316
0
2,728
135
3
242
32
2,250
224
238
150
5,515
0
3,142
220
11
424
69
6,216
207
251
179
1,989
0
2,847
2
6
175
10
3,757
212
229
154
6,194
0
2,536
34
5
242
16
1,242
219
228
146
4,201
0
2,893
21
6
295
36
4,106
216
227
148
5,865
0
2,687
20
11
150
39
1,986
212
217
140
4,936
0
3,109
211
17
424
55
6,089
204
254
181
2,017
0
3,119
200
20
60
0
6,297
208
253
172
2,377
0
2,795
18
3
218
17
3,042
217
233
153
6,073
0
2,555
58
14
285
19
1,231
230
209
109
4,159
0
2,499
42
13
497
-7
953
222
212
123
3,950
0
2,916
105
13
268
89
5,618
242
223
108
4,417
0
2,591
42
12
350
22
1,398
222
213
125
4,224
0
3,083
257
4
510
-25
6,363
209
242
171
2,873
0
2,860
276
33
60
33
5,292
115
226
240
4,759
0
3,008
196
15
228
20
5,535
215
251
166
1,924
0
2,873
130
16
67
16
4,030
244
230
110
6,091
0
2,998
102
12
384
158
6,026
239
225
114
4,121
0
2,658
357
20
300
54
1,597
184
199
149
4,286
0
2,622
95
12
484
39
1,329
239
222
112
4,086
0
2,955
288
21
247
128
5,224
156
233
215
4,968
0
2,837
66
9
201
23
3,655
229
221
124
6,341
0
3,077
129
3
618
43
6,296
225
237
147
3,261
0
2,755
320
4
30
-1
2,890
209
236
165
5,468
0
2,880
86
12
30
3
4,369
237
221
113
5,906
0
2,827
332
15
42
9
3,599
185
221
174
6,067
0
2,518
107
5
360
39
553
229
234
139
3,522
0
End of preview.
StreamArena Logo

A Living Benchmark for Machine Learning on Streaming Data


Python Hugging Face Streaming ML Anomaly Detection Classification Regression Clustering License MIT


πŸ† Live leaderboard: techynilesh.github.io/StreamArena Β· Dataset catalog

StreamArena aggregates datasets for stream learning β€” classification, regression, clustering, and anomaly detection under concept drift β€” into one consistently organized, task-first collection. It plays the same role for streaming/online ML that TabArena plays for tabular ML: a single place to find curated, ready-to-use datasets instead of hunting through individual paper repos.

See the GitHub repo for loaders, examples, and a download.py helper. Datasets were consolidated from several independent research codebases, deduplicated where the same dataset appeared in multiple sources, and reorganized by task. Every dataset is stored as a single unified format β€” CSV β€” chosen because it's what the streaming-ML ecosystem (River's stream.iter_csv, MOA, scikit-multiflow's FileStream) actually consumes row-by-row, unlike batch/columnar formats.

Dataset structure

classification/
β”œβ”€β”€ real/               # real-world streams (electricity, forest cover, airlines, ...)
└── synth/               # synthetic drift generators (SEA, RBF, Hyperplane, Agrawal, Madelon, ...)
regression/
β”œβ”€β”€ real/               # housing, wages, sensor/physical measurements, ...
└── synth/               # Friedman & Hyperplane synthetic generators
clustering/
β”œβ”€β”€ real/               # real-world streams reused from classification
└── synth/               # synthetic drift streams + blobs
anomaly_detection/        # ODDS/ADBench-style outlier detection sets (all real-world)

See DATASETS.md for the full per-dataset table β€” exact instance/feature/class counts computed directly from each file, plus a best-effort source attribution (UCI, OpenML, DELVE, MOA/River generators, ODDS/ADBench, etc.) for every dataset.

All files are .csv. Anomaly-detection files hold feature columns plus a trailing label column; everything else follows the same feature-columns-plus-target convention. Every task except anomaly detection (which is entirely real-world benchmark data) is split into real/ and synth/.

Task Count Notes
Classification 42 files (22 real + 20 synthetic) real/: electricity, forest cover, airlines, poker, weather, KDD-99, insects, Nomao, MNIST, Usenet, Gisette, Dota, Spambase, HAR, etc. synth/: classic drift generators (SEA, RBF, Hyperplane, Agrawal, Madelon)
Regression 30 files (25 real + 5 synthetic) real/: housing (king's county, california, miami, brazilian), wages, sensor/physical (sarcos, naval propulsion, superconductivity, kin8nm), and more. synth/: Friedman & Hyperplane generators
Clustering 13 files (6 real + 7 synthetic) Streaming clustering benchmarks β€” reuses classification drift streams plus a dedicated synthetic blobs set
Anomaly Detection 51 files ODDS/ADBench-style outlier detection collection (annthyroid, mnist, shuttle, satellite, mammography, etc.) β€” all real-world, no real//synth/ split

Usage

pip install huggingface_hub
from huggingface_hub import snapshot_download

path = snapshot_download(repo_id="techynilesh/streamarena", repo_type="dataset")

Or download just one task:

from huggingface_hub import snapshot_download

path = snapshot_download(
    repo_id="techynilesh/streamarena",
    repo_type="dataset",
    allow_patterns=["classification/**"],
)

Then load files directly β€” it's always just a CSV:

import pandas as pd
df = pd.read_csv(f"{path}/classification/real/electricity.csv")

Using it with River or CapyMOA

Since every dataset is plain CSV, it plugs directly into the two most common Python streaming-ML libraries β€” no conversion needed.

# River
import pandas as pd
from river import metrics, stream, tree

path = "classification/real/electricity.csv"
sample = pd.read_csv(path, nrows=100)
target = sample.columns[-1]
# Convert only numeric feature columns to float; categorical/string columns
# (e.g. in adult.csv) pass through as-is β€” River trees handle them natively.
converters = {
    c: float for c in sample.columns[:-1] if pd.api.types.is_numeric_dtype(sample[c])
}

dataset = stream.iter_csv(path, target=target, converters=converters)
model = tree.HoeffdingTreeClassifier()
metric = metrics.Accuracy()

for x, y in dataset:
    y_pred = model.predict_one(x)
    model.learn_one(x, y)
    metric.update(y, y_pred)

print(metric)
# CapyMOA (requires a working JVM β€” Java 11+)
from capymoa.classifier import HoeffdingTree
from capymoa.evaluation import prequential_evaluation
from capymoa.stream import stream_from_file

stream = stream_from_file(
    "classification/real/electricity.csv",
    dataset_name="Electricity",
    class_index=-1,       # StreamArena's convention: label is the trailing column
    target_type="categorical",
)
learner = HoeffdingTree(schema=stream.get_schema())
results = prequential_evaluation(stream, learner)
print("accuracy:", results.cumulative.accuracy())

See examples/river_usage.py and examples/capymoa_usage.py on GitHub for the full runnable scripts.

License

MIT for the aggregation/curation. Individual datasets retain their original licenses/terms from their respective sources β€” check before redistribution.

Citation

If you use StreamArena in your research, please cite it as below:

@misc{verma2026streamarena,
  title  = {StreamArena: A Living Benchmark for Machine Learning on Streaming Data},
  author = {Verma, Nilesh},
  year   = {2026},
  url    = {https://github.com/TechyNilesh/StreamArena}
}

Please also cite the original dataset sources where applicable.

Downloads last month
190