Data Quality (Ratification)¶

All data returned by Aeolus includes a ratification column that indicates the quality or validation status of each measurement. This page explains what each value means and how different data sources handle quality assurance.

Overview¶

The ratification column provides a consistent way to understand data quality across all sources, even though each source has different QA/QC processes:

Value	Meaning	Sources
`Ratified`	Officially validated by network operator	UK regulatory networks
`Provisional`	Awaiting official validation	UK regulatory networks, EPA AirNow
`Validated`	Passed automated QA/QC checks	PurpleAir
`Indicative`	Calibrated low-cost sensor data	AirQo
`Unvalidated`	No formal QA/QC applied	OpenAQ, Breathe London, Sensor.Community, PurpleAir (non-PM)
`Invalid`	Failed QA/QC checks	PurpleAir

Source-Specific Details¶

UK Regulatory Networks (AURN, SAQN, WAQN, NI, AQE)¶

Reference-grade monitors with formal ratification by government bodies.

Value	Meaning
`Ratified`	Data has been reviewed and officially approved
`Provisional`	Recent data awaiting formal review (typically ratified within weeks/months)
`None`	Status not available

Note: Ratified data from UK networks is considered the gold standard for air quality measurements.

OpenAQ¶

Global aggregator of data from many sources with varying quality.

Value	Meaning
`Unvalidated`	All data (OpenAQ aggregates from multiple sources with different QA processes)

Note: OpenAQ data quality depends on the original source. Reference-grade government monitors will be more reliable than low-cost sensors.

Breathe London¶

London sensor network operated by Imperial College London.

Value	Meaning
`Ratified`	Data validated by ERG
`Provisional`	Awaiting validation
`Unvalidated`	Status unknown or not provided

Note: Breathe London uses a mix of reference-grade and indicative monitors.

EPA AirNow¶

US EPA real-time air quality monitoring network.

Value	Meaning
`Provisional`	All data (real-time, preliminary, subject to change)

Note: AirNow data is preliminary and updated hourly. For verified historical US data, use the EPA AQS (Air Quality System) which contains ratified data after a 6+ month delay.

AirQo¶

African air quality network using calibrated low-cost sensors.

Value	Meaning
`Indicative`	All data (machine learning calibrated against reference monitors)

Note: AirQo sensors are calibrated using machine learning models trained on co-located reference monitors. Data is suitable for understanding patterns and trends but may have higher uncertainty than reference-grade measurements.

PurpleAir¶

Global network of dual-channel laser particle counters.

Value	Meaning
`Validated`	Both channels agree within thresholds
`Channel Disagreement`	Both channels valid but disagree beyond thresholds
`Single Channel (A)`	Only channel A had valid data
`Single Channel (B)`	Only channel B had valid data
`Below Detection Limit`	Value below 0.3 µg/m³ (sensor noise floor)
`Sensor Saturation`	Value above 1000 µg/m³ (sensor range exceeded)
`Invalid`	Both channels invalid
`Unvalidated`	Non-PM measurements (temperature, humidity)

QA/QC Thresholds (PM measurements):

Concentration	Agreement Required
< 0.3 µg/m³	Flagged as below detection limit
0.3–100 µg/m³	Channels must agree within ±10 µg/m³
100–1000 µg/m³	Channels must agree within ±10%
> 1000 µg/m³	Flagged as sensor saturation

Note: PurpleAir's dual-channel design allows automated QA/QC. The include_flagged=False parameter can be used to retrieve only validated data.

Sensor.Community¶

Global citizen science network.

Value	Meaning
`Unvalidated`	All data (citizen-installed sensors with no formal QA/QC)

Note: Sensor.Community data is valuable for understanding spatial patterns but has no formal quality control. Sensor placement, maintenance, and calibration vary widely.

Filtering by Quality¶

Get only validated data¶

# Filter to ratified/validated data only
clean_data = data[data['ratification'].isin(['Ratified', 'Validated'])]

PurpleAir: exclude flagged data¶

from aeolus.sources.purpleair import fetch_purpleair_data

# Only retrieve measurements that passed QA/QC
data = fetch_purpleair_data(
    sites=["131075"],
    start_date=start,
    end_date=end,
    include_flagged=False
)

Check quality distribution¶

# See breakdown of quality flags
print(data['ratification'].value_counts())

Recommendations¶

Use Case	Recommended Sources	Minimum Quality
Regulatory compliance	UK networks	`Ratified`
Health studies	UK networks, reference-grade OpenAQ	`Ratified` or `Provisional`
Spatial analysis	Any source	Any (understand limitations)
Trend analysis	Consistent source over time	`Provisional` or better
Real-time alerts	PurpleAir, Sensor.Community	`Validated` or `Unvalidated` with caution