Why Do Air Quality Data Differ Between Portals?

Users often notice that two portals show different AQI values, colors, or categories for the same city.

This does not automatically mean one of them is wrong.

Differences usually result from:

• different AQI methodology
• different monitoring stations
• different update times
• different averaging periods
• different pollutant sets
• different processing and validation

Different AQI methodologies

AirQuality.city uses the US EPA AQI methodology for PM2.5 and PM10.

Other portals may use:

• the European Air Quality Index
• a national scale
• a custom methodology
• different category breakpoints

The same concentration can therefore receive a different number, color, or category name.

AQI is not the same as concentration

PM2.5 and PM10 concentrations are expressed in µg/m³.

That is the measured value.

AQI is an index created by converting concentration according to a selected scale.

If two portals show similar concentrations but different AQI values, methodology is the most likely reason.

If the concentrations also differ, check the station, time, and data source.

Portals may not use the same monitoring station

A city name does not guarantee the same location.

One portal may display:

• a roadside station
• a residential station
• an industrial station
• a background station outside the center
• an average of several stations
• the nearest available station
• the station with the highest value

Differences between micro-locations can be substantial.

Update times differ

One portal may show a value from a few minutes ago, while another shows the last confirmed value from an hour earlier.

When comparing, check:

• the same date
• the same hour
• the same time zone
• the same averaging period

Even a one-hour difference can matter when conditions are changing quickly.

Averaging periods differ

A current reading, one-hour average, eight-hour average, and daily average are not the same.

On AirQuality.city:

• NO₂, O₃, and SO₂ are shown as one-hour values
• CO is shown as an eight-hour rolling average

Directly comparing different averaging periods is not meaningful.

Pollutant sets differ

AirQuality.city bases its overall AQI only on:

• PM2.5
• PM10

Another index may also include:

• NO₂
• O₃
• SO₂
• CO
• other pollutants

Its overall category may therefore be determined by another pollutant.

Operational and validated data

Real-time data are often operational.

An official system may later:

• correct them
• mark them invalid
• remove them
• add missing values

One portal may show an operational value, while another waits for validated data.

SensorBox and reference stations are not the same

SensorBox and similar optical sensors are useful for local insight and trend monitoring.

Their results may be affected by:

• humidity
• fog
• placement
• airflow
• local sources
• sensor condition

Reference stations use different equipment and stricter quality-control procedures.

Different technologies may therefore produce different results, especially in difficult weather conditions.

One portal may aggregate several stations

Some portals display:

• an average of several stations
• the highest value
• the nearest station
• an administratively assigned station
• a modelled value between stations

Without checking the source, it may not be clear whether a number represents one direct measurement or a processed result.

Display and rounding differences

Portals may:

• round to whole numbers
• use a different number of decimal places
• apply filters
• delay display
• handle missing values differently

Small differences may therefore have little practical significance.

How to compare two portals correctly

Check whether they:

1. use the same station
2. show the same time
3. display concentration or AQI
4. use the same AQI methodology
5. use the same averaging period
6. include the same pollutants
7. show operational or validated data
8. show a measurement or an aggregated result

Only then does the comparison become meaningful.

Which portal is correct?

A portal is transparent when it clearly shows:

• the data source
• the station
• the measurement time
• the measurement unit
• the methodology
• the data limitations

Two different results may both be technically correct if they use different methods or sources.