---
license: mit
task_categories:
- table-question-answering
language:
- en
tags:
- legal
pretty_name: nyc-vehicle-crashes
size_categories:
- 1M<n<10M
---
# 🚗 NYC Vehicle Crashes Dataset

This dataset contains detailed records of vehicle collisions in New York City, made available for public use by the NYC Open Data initiative. The source CSV file is `NYC_Vehicle_Collisions.csv`, and this dataset is now accessible via the Hugging Face Datasets Hub.

## 📁 Dataset Structure

Each row represents a reported vehicle crash incident and includes fields such as:

- `CRASH_DATE`
- `CRASH_TIME`
- `BOROUGH`
- `ZIP CODE`
- `LATITUDE`
- `LONGITUDE`
- `ON STREET NAME`
- `CROSS STREET NAME`
- `OFF STREET NAME`
- `NUMBER OF PERSONS INJURED`
- `NUMBER OF PERSONS KILLED`
- `NUMBER OF PEDESTRIANS INJURED`
- `NUMBER OF PEDESTRIANS KILLED`
- `NUMBER OF CYCLIST INJURED`
- `NUMBER OF CYCLIST KILLED`
- `NUMBER OF MOTORIST INJURED`
- `NUMBER OF MOTORIST KILLED`
- `CONTRIBUTING FACTOR VEHICLE 1`
- `CONTRIBUTING FACTOR VEHICLE 2`
- `CONTRIBUTING FACTOR VEHICLE 3`
- `CONTRIBUTING FACTOR VEHICLE 4`
- `CONTRIBUTING FACTOR VEHICLE 5`	
- `COLLISION_ID`
- `VEHICLE TYPE CODE 1`
- `VEHICLE TYPE CODE 2`	
- `VEHICLE TYPE CODE 3`
- `VEHICLE TYPE CODE 4`
- `VEHICLE TYPE CODE 5`


> This dataset is intended for analysis, visualization, machine learning, or policy research related to road safety and traffic incidents.

---

## 📥 Getting Started

### Step 1: Install `datasets` library

If you haven’t already:

```bash
pip install datasets
```

### Step 2: Load the Dataset from Hugging Face

Use the following Python code to load the dataset:

```python
from datasets import load_dataset

dataset = load_dataset("langutang/nyc-vehicle-crashes")

# View first row
print(dataset['train'][0])
```

> Replace `"your-username"` with your actual Hugging Face username or organization name if applicable.

### Optional: Load Specific Splits (if defined)

If the dataset is split into `train`, `test`, or other subsets, you can access them like:

```python
train_data = dataset['train']
test_data = dataset['test']
```

---

## 📊 Example Use Case

```python
import pandas as pd

# Convert to pandas DataFrame
df = dataset['train'].to_pandas()

# Analyze crashes by borough
print(df['BOROUGH'].value_counts())
```

---

## 📎 Citation & Source

Original data provided by: [NYC Open Data – Motor Vehicle Collisions](https://data.cityofnewyork.us/Transportation/Motor-Vehicle-Collisions-Vehicles/bm4k-52h4)

If you use this dataset in your research or application, please cite the source and the Hugging Face repository.

---

## 🤝 Contributions

If you'd like to improve or extend this dataset (e.g., preprocessing, cleaning, or labeling), feel free to open a PR or discuss it in the [Discussions](https://huggingface.co/datasets/langutang/nyc-vehicle-crashes/discussions) tab.

---

## 📬 Contact

Questions or suggestions? Reach out on [Hugging Face](https://huggingface.co/langutang) or submit an issue in the repo.