meganariley/daily-aqi

Daily AQI — US Air Quality

Sharing datasets helps agents analyze them — giving everyone the ability to make sense of complex data.

Official US EPA air quality data updated daily. All 6 NAAQS criteria pollutants, hourly granularity, ~1,500 monitoring stations across the United States.

Explore it interactively in the Daily AQI Space.

Dataset Structure

readings.parquet — hourly readings

One row per monitoring station per hour per pollutant.

Column	Type	Description
timestamp	datetime	Local measurement timestamp (hourly)
pollutant	str	One of: pm25, pm10, ozone, so2, co, no2
value	float32	Raw concentration measurement
unit	str	Unit of measure (e.g. Parts per billion, µg/m³)
aqi	Int16	Calculated AQI for this reading (EPA standard breakpoints)
state_name	str	US state name
county_name	str	County name
state_code	Int16	FIPS state code
county_code	Int16	FIPS county code
site_num	Int32	EPA monitoring site number
latitude	float32	Station latitude
longitude	float32	Station longitude
mdl	float	Method detection limit
datum	str	Coordinate datum (WGS84, NAD83, etc.)

daily_summary.parquet — daily aggregates

Daily mean/max/min AQI per state, county, and pollutant. Lighter weight — use this for visualisation and trend analysis.

Column	Type	Description
date	datetime	Date
pollutant	str	Pollutant code
state_name	str	US state name
county_name	str	County name
aqi_mean	float32	Mean AQI across all readings that day
aqi_max	Int16	Max AQI reading
aqi_min	Int16	Min AQI reading
value_mean	float32	Mean raw concentration
reading_count	Int32	Number of hourly readings

AQI Reference

AQI Range	Category
0–50	Good
51–100	Moderate
101–150	Unhealthy for Sensitive Groups
151–200	Unhealthy
201–300	Very Unhealthy
301–500	Hazardous

Usage

import pandas as pd
from huggingface_hub import hf_hub_download

# Daily summary (recommended for analysis and visualisation)
summary = pd.read_parquet(
    hf_hub_download("meganariley/daily-aqi", "daily_summary.parquet", repo_type="dataset")
)

# Full hourly data (for ML training)
readings = pd.read_parquet(
    hf_hub_download("meganariley/daily-aqi", "readings.parquet", repo_type="dataset")
)

# Example: worst air quality days in California
ca = summary[(summary["state_name"] == "California") & (summary["pollutant"] == "pm25")]
print(ca.sort_values("aqi_max", ascending=False).head(10))

Data Source & Updates

Data sourced from the EPA Air Quality System (AQS) — the official US government air quality monitoring network.

Updated daily via automated pipeline. New readings are appended using Xet block-level deduplication, meaning only changed data blocks are stored on each update.

Licensed under Open Data Commons Attribution License (ODC-By).