claritystorm/osha-workplace-injuries

---
license: other
license_name: public-domain
task_categories:
- tabular-classification
- tabular-regression
tags:
- workplace-safety
- osha
- insurance
- esg
- compliance
- united-states
pretty_name: OSHA Workplace Injuries & Illness 2016-2023
size_categories:
- 1M<n<10M
---

# OSHA Workplace Injuries & Illness 2016–2023

Every establishment-level workplace injury and illness record submitted to the OSHA Injury Tracking Application (ITA) since 2016.
**2.38 million records** across 8 survey years covering 600K+ unique establishments — with NAICS codes, DART/TCIR rates, and detailed injury/illness breakdowns.
The only public dataset linking individual establishment safety performance to industry benchmarks.

| 📊 Records | 📅 Coverage | 🏷️ License | 🔄 Updated |
|-----------|-------------|-----------|-----------|
| 2.38M+ records | 2016–2023 (8 years) | Public Domain | Annual |

**This repo contains a free 1,000-row sample.**
Full dataset (CSV + Parquet) → **[claritystorm.com/datasets/osha-injuries](https://claritystorm.com/datasets/osha-injuries)**

---

## Quick Start

```python
from datasets import load_dataset
import pandas as pd

# Load the 1,000-row sample
ds = load_dataset("claritystorm/osha-workplace-injuries")
df = ds["train"].to_pandas()

# Total workplace deaths by year
print(df.groupby("survey_year")["total_deaths"].sum())

# Industries with highest average DART rate
print(df.groupby("industry_description")["dart_rate"].mean()
.sort_values(ascending=False).head(10))

# Establishments with zero injuries (benchmark group)
safe_pct = df["no_injuries_illnesses"].mean() * 100
print(f"Establishments with zero injuries: {safe_pct:.1f}%")

# Size class distribution
print(df["size_class"].value_counts())
```

## Use Cases

- **Workplace safety risk scoring** — predict DART/TCIR rates from NAICS code, size class, and historical performance
- **ESG & responsible investing** — screen company supply chains and subsidiaries for OSHA safety performance
- **Insurance underwriting** — establishment-level injury rates for workers' compensation risk modeling
- **OSHA compliance benchmarking** — compare an establishment's safety record to industry averages
- **Industry safety trend analysis** — 8-year panel data tracks safety improvements and deterioration by sector
- **Human capital ML** — injury/illness rates as a feature for company quality and labor conditions scoring

## Schema (selected fields)

| Field | Type | Description |
|-------|------|-------------|
| survey_year | int | OSHA reporting year (2016–2023) |
| estab_name | string | Establishment name |
| company_name | string | Parent company name |
| state | string | US state (2-letter code) |
| naics_code | string | 6-digit NAICS code |
| industry_description | string | Industry description |
| size_class | string | Establishment size class |
| annual_average_employees | int | Annual average employee count |
| total_hours_worked | int | Total hours worked |
| total_deaths | int | Total workplace deaths |
| total_dafw_cases | int | Days Away From Work cases |
| total_injuries | int | Total injuries |
| total_resp_conditions | int | Respiratory condition cases |
| dart_rate | float | DART rate per 100 FTE (derived) |
| tcir_rate | float | TCIR rate per 100 FTE (derived) |

## ⬇️ Get the Full Dataset

| Tier | Price | Includes |
|------|-------|----------|
| Sample | Free | 1,000 rows, Public Domain (this repo) |
| Complete | $79 | Full 2.38M+ rows, CSV + Parquet, commercial license |
| Annual | $149/yr | Complete + annual updates |

👉 **[Purchase at claritystorm.com/datasets/osha-injuries](https://claritystorm.com/datasets/osha-injuries)**

## Source

US Occupational Safety and Health Administration (OSHA), Injury Tracking Application (ITA).
OSHA ITA data is a US federal government work in the **public domain** (17 U.S.C. 105).

Processed and structured by [ClarityStorm Data](https://claritystorm.com).