Posted by:

David G

David G, Chief Explorer

In this post I will show you how to find the air quality at the time and location a photo was taken.

Did you read my post in 2020; Measuring the Air Quality of Your Trek, in which I used a Atmotube PRO to measure air quality.

The Atmotube PRO detected PM1, PM2.5, and PM10 pollutants, like dust, pollen, soot, and mold, plus a wide range of Volatile Organic Compounds (VOCs)

Based on the feedback from readers who are far more knowledgeable in the field, it became clear the AtmoTube Pro works best when still and placed indoors. For measuring air quality on the move in an accurate way would be impossible. Even more so, it would also be unrealistic for all Explorer users to purchase an Atmotube PRO.

So I went back to the drawing board.

It turns out there are a few data service that offer historical air quality data.

Here are the two most promising options I found during my research;

I then wanted to put them to a test.

To do this, I used a photo taken in a fairly remote location (Helvellyn, Lake District, UK) because I wanted to test how measurements were reported when far from a weather station.

GSAL1592 Helvellyn

Here is the metadata in the photo needed to make the requests to the APIs.

exiftool -GPSLatitude -GPSLongitude -GPSDateTime -n GSAL1592.JPG

Exiftool by default prints data as Degrees, Minutes, and Seconds (DMS). Using the -n flag turns DMS into Decimal Degrees (DD), required for both the APIs.

GPS Latitude                    : 54.5255907
GPS Longitude                   : -3.003587
GPS Date/Time                   : 2022:07:13 12:13:07Z

OpenWeather Air Pollution API

The request to OpenWeather Historic Air Pollution API takes the following structure:

http://api.openweathermap.org/data/2.5/air_pollution/history?lat={lat}&lon={lon}&start={start}&end={end}&appid=ba9eef2e11e65a09842f26b115567593

The {start} and {end} times need to be reported as epoch time. There are lots of epoch convertors online, for example epochconverter.com.

The epoch time for this photo is: 1657714387 However, the OpenWeather Historic Air Pollution API reports data hourly and only returns measurements between {start} and {end} defined in the request.

Therefore, to build the request I snap the time to the nearest hour either side of the photo time (2022:07:13 12:13:07)

  • start = 2022:07:13 12:00:01Z = 1657713601
  • end = 2022:07:13 13:00:00Z = 1657717200

So for this photo the request is as follows;

http://api.openweathermap.org/data/2.5/air_pollution/history?lat=54.5255907&lon=-3.003587&start=1657713601&end=1657717200&appid=REDACTED

Which returns one result (at time 2022:07:13 13:00:00 / "dt": 1657717200):

{
  "coord": {
    "lon": -3.0036,
    "lat": 54.5256
  },
  "list": [
    {
      "main": {
        "aqi": 1
      },
      "components": {
        "co": 153.54,
        "no": 0.09,
        "no2": 0.4,
        "o3": 52.21,
        "so2": 0.54,
        "pm2_5": 1.91,
        "pm10": 3.56,
        "nh3": 0.8
      },
      "dt": 1657717200
    }
  ]
}

All the fields are described in the OpenWeather documentation here.

The Air Quality Index (list.main.aqi) reported by the OpenWeather API maps to the Common Air Quality Index (CAQI) values as follows;

  • 1 (Good) = 0–25 (Very low)
  • 2 (Fair) = 25–50 (Low)
  • 3 (Moderate) = 50–75 (Medium)
  • 4 (Poor) = 75–100 (High)
  • 5 (Very Poor) = >100 (Very high)

Besides the aggregated Air Quality Index, the API returns data about the following polluting gases;

  1. Carbon monoxide (CO) (list.components.co),
  2. Nitrogen monoxide (NO) (list.components.no),
  3. Nitrogen dioxide (NO2) (list.components.no2),
  4. Ozone (O3) (list.components.o3),
  5. Sulphur dioxide (SO2) (list.components.so2),
  6. Ammonia (NH3) (list.components.nh3),
  7. Particulates PM2.5 (list.components.pm2_5),
  8. Particulates PM10 (list.components.pm10)

The biggest limitations I identified with this API were;

  • Historical data is accessible from 27th November 2020.
  • The coord value returned is the photo location. There is no information about the weather station that recorded the measurements (which could be very far away).

WeatherBit Historical Air Quality API

The request to WeatherBit Historical Air Quality API takes the following structure;

http://api.weatherbit.io/v2.0/history/airquality?lat={lat}&lon={lon}&start_date={start}&end_date={end}&tz=local&key={api_key}

The data is reported hourly, and the {start} and {end} dates must be different. In my example I use; 2022-07-13 and 2022-07-14. Therefore, the returned data will always cover 24 points for each hour of the day.

For the example photo I would make the request;

http://api.weatherbit.io/v2.0/history/airquality?lat=54.5255907&lon=-3.003587&start_date=2022-07-13&end_date=2022-07-14&key=REDACTED

As noted, this returns hourly measurements for the day (2022-07-13). Below I have modified the response to only include the response for the same time period shown in the OpenWeather API response (1657717200 = Wednesday, 13 July 2022 13:00:00Z);

{
  "lat":54.5255907,
    "data":[
      {
        "aqi": 26,
        "pm10": 3.28,
        "pm25": 1.57,
        "o3": 56.78,
        "timestamp_local": "2022-07-13T14:00:00",
        "so2": 0.45,
        "no2": 0.33,
        "timestamp_utc": "2022-07-13T13:00:00",
        "datetime": "2022-07-13:13",
        "co": 155.11,
        "ts": 1657717200
      }
    ],
  "timezone":"Europe\/London",
  "country_code":"GB",
  "lon":-3.003587,
  "city_name":"Ambleside",
  "state_code":"ENG"
}

All the fields are described in the Weatherbit documentation here.

The response includes the Air Quality Index (data.aqi) using the CAQI values (so the data does not need to be normalised like in the case of the OpenWeather API).

Besides the aggregated Air Quality Index, the API returns data about the following polluting gases;

  1. o3 (data.o3)
  2. so2 (data.so2)
  3. no2 (data.no2)
  4. co (data.co)
  5. pm25 (data.pm25)
  6. pm10 (data.pm10)

The biggest limitations I identified with this API were;

  • Data is only available going back to January 13th, 2022 at this time. Much less than the OpenWeather API.
  • The response does not include Ammonia (NH3) and Nitrogen monoxide (NO) reported by OpenWeather API (although for most cases, I am not sure how valuable these reading will be).
  • The lat and lon values returned is the photo location. There is no information about the weather station that recorded the measurements (which could be very far away).

My verdict

I chose the OpenWeather API for Explorer.

The data in each response is similar, with the OpenWeather API containing a few more measurements in the response (that I don’t really need).

Though ultimately it came down to cost. I could not justify (or afford) the monthly cost of the WeatherBit Historical Air Quality API with no additional features.

Explorer logic for air quality lookup

OpenWeather’s free plan has a few key restrictions – limits 60 calls/minute and 1,000,000 calls/month.

Keep in mind, Explorer Sequences can often contain 1000 or more photos. If Explorer was to lookup air quality against each image, the calls could quickly add up.

I thought looking up air quality for each photo was overkill. Given sequences are usually no more than twenty minutes long and photos in them usually no more than 5 seconds apart I decided to assign air quality at sequence level.

Explorer does this by taking the metadata from the first image in the sequence and making a request to the OpenWeather Air Pollution API. As noted earlier, the OpenWeather API requires a time range to be supplied in the request. In the same way I did before by snapping the time to the closest hours either side of the photo time.

It is not perfect, but given air quality does not usually change particularly quickly, having a margin of error of +/- 1 hour was deemed acceptable for how air quality is used in Explorer.

Note, if you are reading this trying to debug why air quality data is not showing against your sequence in Explorer it is probably because it was taken before the 27th November 2020 (the OpenWeather API data does not report Air Quality data before this point).

If you can recommend any other air quality data sources, please do share them with me – I can add them to this post for others to benefit from.

Similar Posts You Will Enjoy Reading...

Never miss an update


Sign up to receive new articles in your inbox as they published.

Discuss this post


Signals Corps Slack
JOIN US