Datasets:

origindatalab
/

bangladesh-rickshaw-traffic-free-sample-v1

Error code:   StreamingRowsError
Exception:    CastError
Message:      Couldn't cast
release_pipeline_version: string
clip_count: int64
metadata_count: int64
privacy_status: string
source_of_truth: string
generated_by: string
render_mode: string
summary: struct<clip_count: int64, unique_sources: int64, unique_sessions: int64, unique_collectors: int64, u (... 390 chars omitted)
  child 0, clip_count: int64
  child 1, unique_sources: int64
  child 2, unique_sessions: int64
  child 3, unique_collectors: int64
  child 4, unique_capture_dates: int64
  child 5, road_context_distribution: struct<urban-road: int64, market-zone: int64>
      child 0, urban-road: int64
      child 1, market-zone: int64
  child 6, mobility_mix_distribution: struct<mixed-mobility: int64, pedestrian-dominant: int64>
      child 0, mixed-mobility: int64
      child 1, pedestrian-dominant: int64
  child 7, avg_buyer_value_score: double
  child 8, avg_commercial_use_readiness_score: double
  child 9, avg_metadata_quality_score: double
  child 10, scene_distribution: struct<intersection: int64, market: int64>
      child 0, intersection: int64
      child 1, market: int64
  child 11, unique_scene_labels: int64
deliverables: struct<clips_blurred: string, metadata: string, sample_index: string, buyer_sample_index: string, ma (... 59 chars omitted)
  child 0, clips_blurred: string
  child 1, metadata: string
  child 2, sample_index: string
  child 3, buyer_sample_index: string
  child 4, manifest: string
  child 5, release_context: string
  child 6, checksums: string
scene_distribution: struct<i
...
, avg_metadata_quality: double
unique_scene_labels: int64
human_gt_included: bool
files: struct<sample_index: string, buyer_sample_index: string, manifest: string, release_context: string,  (... 18 chars omitted)
  child 0, sample_index: string
  child 1, buyer_sample_index: string
  child 2, manifest: string
  child 3, release_context: string
  child 4, checksums: string
validation: struct<forbidden_terms_checked: bool, validators_read_only: bool, human_document_edit_required: bool (... 1 chars omitted)
  child 0, forbidden_terms_checked: bool
  child 1, validators_read_only: bool
  child 2, human_document_edit_required: bool
catalog_version: string
enterprise_buyer_summary: struct<why_buy_this_dataset: list<item: string>, recommended_use_cases: list<item: string>, limitati (... 235 chars omitted)
  child 0, why_buy_this_dataset: list<item: string>
      child 0, item: string
  child 1, recommended_use_cases: list<item: string>
      child 0, item: string
  child 2, limitations: list<item: string>
      child 0, item: string
  child 3, capture_method: string
  child 4, privacy_method: string
  child 5, geographic_coverage: string
  child 6, dataset_statistics: struct<clips: int64, metadata_files: int64, human_gt_images: int64, human_gt_labels: int64, indexed_ (... 12 chars omitted)
      child 0, clips: int64
      child 1, metadata_files: int64
      child 2, human_gt_images: int64
      child 3, human_gt_labels: int64
      child 4, indexed_rows: int64
product_type: string
to
{'catalog_version': Value('string'), 'product_type': Value('string'), 'source_of_truth': Value('string'), 'privacy_status': Value('string'), 'release_pipeline_version': Value('string'), 'clip_count': Value('int64'), 'metadata_count': Value('int64'), 'human_gt_included': Value('bool'), 'summary': {'clip_count': Value('int64'), 'unique_sources': Value('int64'), 'unique_sessions': Value('int64'), 'unique_collectors': Value('int64'), 'unique_capture_dates': Value('int64'), 'road_context_distribution': {'urban-road': Value('int64'), 'market-zone': Value('int64')}, 'mobility_mix_distribution': {'mixed-mobility': Value('int64'), 'pedestrian-dominant': Value('int64')}, 'avg_buyer_value_score': Value('float64'), 'avg_commercial_use_readiness_score': Value('float64'), 'avg_metadata_quality_score': Value('float64'), 'scene_distribution': {'intersection': Value('int64'), 'market': Value('int64')}, 'unique_scene_labels': Value('int64')}, 'files': {'sample_index': Value('string'), 'buyer_sample_index': Value('string'), 'manifest': Value('string'), 'release_context': Value('string'), 'checksums': Value('string')}, 'validation': {'forbidden_terms_checked': Value('bool'), 'validators_read_only': Value('bool'), 'human_document_edit_required': Value('bool')}, 'scene_distribution': {'intersection': Value('int64'), 'market': Value('int64')}, 'enterprise_buyer_summary': {'why_buy_this_dataset': List(Value('string')), 'recommended_use_cases': List(Value('string')), 'limitations': List(Value('string')), 'capture_method': Value('string'), 'privacy_method': Value('string'), 'geographic_coverage': Value('string'), 'dataset_statistics': {'clips': Value('int64'), 'metadata_files': Value('int64'), 'human_gt_images': Value('int64'), 'human_gt_labels': Value('int64'), 'indexed_rows': Value('int64')}}, 'enterprise_truth': {'clips': Value('int64'), 'metadata': Value('int64'), 'human_gt_images': Value('int64'), 'human_gt_labels': Value('int64'), 'unique_sessions': Value('int64'), 'unique_sources': Value('int64'), 'unique_capture_dates': Value('int64'), 'avg_buyer_value': Value('float64'), 'avg_commercial_readiness': Value('float64'), 'avg_metadata_quality': Value('float64')}, 'unique_scene_labels': Value('int64')}
because column names don't match
Traceback:    Traceback (most recent call last):
                File "/src/services/worker/src/worker/utils.py", line 147, in get_rows_or_raise
                  return get_rows(
                      dataset=dataset,
                  ...<4 lines>...
                      column_names=column_names,
                  )
                File "/src/libs/libcommon/src/libcommon/utils.py", line 272, in decorator
                  return func(*args, **kwargs)
                File "/src/services/worker/src/worker/utils.py", line 127, in get_rows
                  rows_plus_one = list(itertools.islice(safe_iter(ds, dataset=dataset), rows_max_number + 1))
                File "/src/services/worker/src/worker/utils.py", line 478, in safe_iter
                  yield from ds.decode(False) if ds.features else ds
                File "/usr/local/lib/python3.14/site-packages/datasets/iterable_dataset.py", line 2818, in __iter__
                  for key, example in ex_iterable:
                                      ^^^^^^^^^^^
                File "/usr/local/lib/python3.14/site-packages/datasets/iterable_dataset.py", line 2355, in __iter__
                  for key, pa_table in self._iter_arrow():
                                       ~~~~~~~~~~~~~~~~^^
                File "/usr/local/lib/python3.14/site-packages/datasets/iterable_dataset.py", line 2380, in _iter_arrow
                  for key, pa_table in self.ex_iterable._iter_arrow():
                                       ~~~~~~~~~~~~~~~~~~~~~~~~~~~~^^
                File "/usr/local/lib/python3.14/site-packages/datasets/iterable_dataset.py", line 536, in _iter_arrow
                  for key, pa_table in iterator:
                                       ^^^^^^^^
                File "/usr/local/lib/python3.14/site-packages/datasets/iterable_dataset.py", line 419, in _iter_arrow
                  for key, pa_table in self.generate_tables_fn(**gen_kwags):
                                       ~~~~~~~~~~~~~~~~~~~~~~~^^^^^^^^^^^^^
                File "/usr/local/lib/python3.14/site-packages/datasets/packaged_modules/json/json.py", line 343, in _generate_tables
                  self._cast_table(pa_table, json_field_paths=json_field_paths),
                  ~~~~~~~~~~~~~~~~^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
                File "/usr/local/lib/python3.14/site-packages/datasets/packaged_modules/json/json.py", line 132, in _cast_table
                  pa_table = table_cast(pa_table, self.info.features.arrow_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 2297, in cast_table_to_schema
                  raise CastError(
                  ...<3 lines>...
                  )
              datasets.table.CastError: Couldn't cast
              release_pipeline_version: string
              clip_count: int64
              metadata_count: int64
              privacy_status: string
              source_of_truth: string
              generated_by: string
              render_mode: string
              summary: struct<clip_count: int64, unique_sources: int64, unique_sessions: int64, unique_collectors: int64, u (... 390 chars omitted)
                child 0, clip_count: int64
                child 1, unique_sources: int64
                child 2, unique_sessions: int64
                child 3, unique_collectors: int64
                child 4, unique_capture_dates: int64
                child 5, road_context_distribution: struct<urban-road: int64, market-zone: int64>
                    child 0, urban-road: int64
                    child 1, market-zone: int64
                child 6, mobility_mix_distribution: struct<mixed-mobility: int64, pedestrian-dominant: int64>
                    child 0, mixed-mobility: int64
                    child 1, pedestrian-dominant: int64
                child 7, avg_buyer_value_score: double
                child 8, avg_commercial_use_readiness_score: double
                child 9, avg_metadata_quality_score: double
                child 10, scene_distribution: struct<intersection: int64, market: int64>
                    child 0, intersection: int64
                    child 1, market: int64
                child 11, unique_scene_labels: int64
              deliverables: struct<clips_blurred: string, metadata: string, sample_index: string, buyer_sample_index: string, ma (... 59 chars omitted)
                child 0, clips_blurred: string
                child 1, metadata: string
                child 2, sample_index: string
                child 3, buyer_sample_index: string
                child 4, manifest: string
                child 5, release_context: string
                child 6, checksums: string
              scene_distribution: struct<i
              ...
              , avg_metadata_quality: double
              unique_scene_labels: int64
              human_gt_included: bool
              files: struct<sample_index: string, buyer_sample_index: string, manifest: string, release_context: string,  (... 18 chars omitted)
                child 0, sample_index: string
                child 1, buyer_sample_index: string
                child 2, manifest: string
                child 3, release_context: string
                child 4, checksums: string
              validation: struct<forbidden_terms_checked: bool, validators_read_only: bool, human_document_edit_required: bool (... 1 chars omitted)
                child 0, forbidden_terms_checked: bool
                child 1, validators_read_only: bool
                child 2, human_document_edit_required: bool
              catalog_version: string
              enterprise_buyer_summary: struct<why_buy_this_dataset: list<item: string>, recommended_use_cases: list<item: string>, limitati (... 235 chars omitted)
                child 0, why_buy_this_dataset: list<item: string>
                    child 0, item: string
                child 1, recommended_use_cases: list<item: string>
                    child 0, item: string
                child 2, limitations: list<item: string>
                    child 0, item: string
                child 3, capture_method: string
                child 4, privacy_method: string
                child 5, geographic_coverage: string
                child 6, dataset_statistics: struct<clips: int64, metadata_files: int64, human_gt_images: int64, human_gt_labels: int64, indexed_ (... 12 chars omitted)
                    child 0, clips: int64
                    child 1, metadata_files: int64
                    child 2, human_gt_images: int64
                    child 3, human_gt_labels: int64
                    child 4, indexed_rows: int64
              product_type: string
              to
              {'catalog_version': Value('string'), 'product_type': Value('string'), 'source_of_truth': Value('string'), 'privacy_status': Value('string'), 'release_pipeline_version': Value('string'), 'clip_count': Value('int64'), 'metadata_count': Value('int64'), 'human_gt_included': Value('bool'), 'summary': {'clip_count': Value('int64'), 'unique_sources': Value('int64'), 'unique_sessions': Value('int64'), 'unique_collectors': Value('int64'), 'unique_capture_dates': Value('int64'), 'road_context_distribution': {'urban-road': Value('int64'), 'market-zone': Value('int64')}, 'mobility_mix_distribution': {'mixed-mobility': Value('int64'), 'pedestrian-dominant': Value('int64')}, 'avg_buyer_value_score': Value('float64'), 'avg_commercial_use_readiness_score': Value('float64'), 'avg_metadata_quality_score': Value('float64'), 'scene_distribution': {'intersection': Value('int64'), 'market': Value('int64')}, 'unique_scene_labels': Value('int64')}, 'files': {'sample_index': Value('string'), 'buyer_sample_index': Value('string'), 'manifest': Value('string'), 'release_context': Value('string'), 'checksums': Value('string')}, 'validation': {'forbidden_terms_checked': Value('bool'), 'validators_read_only': Value('bool'), 'human_document_edit_required': Value('bool')}, 'scene_distribution': {'intersection': Value('int64'), 'market': Value('int64')}, 'enterprise_buyer_summary': {'why_buy_this_dataset': List(Value('string')), 'recommended_use_cases': List(Value('string')), 'limitations': List(Value('string')), 'capture_method': Value('string'), 'privacy_method': Value('string'), 'geographic_coverage': Value('string'), 'dataset_statistics': {'clips': Value('int64'), 'metadata_files': Value('int64'), 'human_gt_images': Value('int64'), 'human_gt_labels': Value('int64'), 'indexed_rows': Value('int64')}}, 'enterprise_truth': {'clips': Value('int64'), 'metadata': Value('int64'), 'human_gt_images': Value('int64'), 'human_gt_labels': Value('int64'), 'unique_sessions': Value('int64'), 'unique_sources': Value('int64'), 'unique_capture_dates': Value('int64'), 'avg_buyer_value': Value('float64'), 'avg_commercial_readiness': Value('float64'), 'avg_metadata_quality': Value('float64')}, 'unique_scene_labels': Value('int64')}
              because column names don't match

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.

Bangladesh Urban Traffic Dataset (Free Sample) for Physical AI, Robotics and ADAS

Privacy-preserved urban traffic dataset from Bangladesh for Physical AI, Robotics, ADAS, autonomous driving and computer vision research.

Free evaluation sample of a Bangladesh urban traffic video dataset for Physical AI, Robotics, ADAS, autonomous driving, object detection and computer vision research.

Collected in real urban road environments across Dhaka, Bangladesh.

This dataset contains privacy-preserved traffic videos, metadata and Human GT preview sheets collected in real urban road environments.

Human GT Preview

Representative Human GT preview sheets generated from the commercial dataset to demonstrate annotation quality and scene diversity.

Human GT Preview Sheet 1

Human GT Preview Sheet 2

Note

The Human GT Preview Sheets demonstrate annotation quality only. Original annotation files are not included in this free sample.

Included

12 curated traffic clips
Human GT Preview sheets
Metadata for every clip
Privacy processed videos
Commercial dataset available separately
The commercial edition includes substantially more video clips, metadata, annotation assets and additional documentation than this free evaluation sample.

Origin Data Lab Enterprise Release

Dataset Summary

Company: Origin Data Lab Co., Ltd.
Contact: contact@origindatalab.io
Website: https://origindatalab.io
Location: Dhaka, Bangladesh
Clip count: 12
Golden source: metadata/*.json
Release pipeline: v2
Privacy status: PRIVACY_REDUCED
Deliverable video folder: clips_blurred/

Buyer Highlights

Urban mixed-traffic scenes from Bangladesh
Rickshaw / motorcycle / pedestrian / vehicle-rich road environments
Privacy-reduced release clips
Metadata-first deterministic release generation
SHA256SUMS.txt generated after all release artifacts are rendered

Diversity Summary

Unique sources: 12
Unique sessions: 12
Unique collectors: 7
Unique capture dates: 7
Unique scene labels: 2
Average buyer value score: 78.75
Average commercial use readiness score: 70.75
Average metadata quality score: 70.75

Scene Distribution

intersection: 10
market: 2

Road Context Distribution

market-zone: 9
urban-road: 3

Mobility Mix Distribution

mixed-mobility: 10
pedestrian-dominant: 2

clips_blurred/
metadata/
human_gt_preview/
sample_index.csv
buyer_sample_index.csv
dataset_manifest.json
release_context.json
SHA256SUMS.txt
DATASET_CARD.md
QUALITY_ASSURANCE.md
PRIVACY_REPORT.md
RELEASE_CERTIFICATE.md
LICENSE.md

Pipeline Rule

All buyer-facing JSON and Markdown files are regenerated from release indexes and metadata. Scene distribution is rendered from sample_index.csv to match buyer-facing CSV taxonomy. Manual post-processing is not required.

Intended Use

This release is designed for evaluation, research, dataset benchmarking, computer vision prototyping, annotation planning, and buyer-side diligence for emerging-market urban mobility datasets.

Recommended Use Cases

Urban traffic perception research
Rickshaw, pedestrian, motorcycle, and mixed-mobility scene analysis
Object detection and tracking dataset evaluation
Annotation workflow planning
Edge-case discovery for dense, unstructured road environments
Multimodal metadata quality review

Limitations

This dataset is not a certified safety validation dataset.
Production deployment, regulated vehicle use, or high-risk automation requires separate commercial review and licensing.
Scene labels and quality scores are metadata aids, not legal or safety certifications.
GPS, motion, and sensor-derived fields may vary by device quality, collection conditions, and signal availability.

Sensor Notes

Metadata may include capture timing, motion, GPS-derived movement indicators, quality scores, scene tags, and buyer-facing readiness fields when available. Sensor-derived values should be treated as supporting metadata for filtering and analysis rather than ground-truth physical measurements.

Capture Method

Clips were selected from field-captured urban mobility video sources, privacy processed, indexed, and packaged through the Origin Data Lab release pipeline. Release documents and CSV indexes are generated from metadata as the source of truth.

Geographic Coverage

Primary coverage: Dhaka, Bangladesh urban road and market-zone mobility scenes.

Privacy Method

Faces and visible license plates are processed through the privacy reduction pipeline. Buyer-facing files avoid exposing raw source video keys and use source_asset_ref for safe asset referencing.

Dataset Statistics

Clips: 12
Metadata files: 12
Human GT preview sheets: 2
Human GT annotations: Not included
Indexed rows: 12
Privacy processed video folder: clips_blurred/

Downloads last month: 129

Total file size:

263 MB