Challenge 16 - Air quality dashboard

[RubenRT7]

Challenge 16 - Air quality dashboard

Stream 1 - Data Visualization and visual narratives for Earth Sciences applications

Goal

Develop an online dashboard for displaying and monitoring air quality forecasts for cities around the world.

Mentors and skills

• Mentors:
  • Mark Parrington, Miha Razinger, Johannes Flemming (all ECMWF)
• Skills required:
  • Python programming (xarray, numpy, pandas, matplotlib, cartopy, seaborn)
  • Processing and visualization of scientific data (netCDF, GRIB…)
  • Front end data visualization

---

Challenge description

The Copernicus Atmosphere Monitoring Service (CAMS) provides forecasts of air quality around the world. CAMS products have a wide reach to the general public through smartphone apps and, notably, through daily broadcasts of air quality index (AQI) forecasts for different cities by CNN International. Routine monitoring and evaluation of the AQI forecasts is required in order to understand short-term changes in air quality, and to be able to respond to any issues arising in the forecast compared to local measurements.

AQI forecasts provided by CAMS are based on a definition from the European Environment Agency which combines the relative concentrations of atmospheric pollutants: ozone, nitrogen dioxide, sulphur dioxide, and fine and coarse particulate matter (PM2.5 and PM10). Monitoring each component along with the combined AQI is therefore important in providing some context.

The aim of this challenge is to develop a dashboard to evaluate AQI, and the key atmospheric pollutants, from the CAMS global forecasts with near-real-time measurements available from different networks (such as OpenAQ), and to visualize the results for the cities covered by CNN.

The project will be divided into the following stages:

1. Data collection: Collection of pre-processed AQI forecast data as provided to CNN, available from ECMWF, collection of air pollutant forecast data at the same locations, collection of in situ air quality measurements (some data available from ECMWF, or via OpenAQ API for additional locations).
2. Data evaluation: Compare forecast AQI vs values calculated from in situ measurements and also individual atmospheric pollutants. An alert system based on divergences of the forecast vs measured values to highlight any potential issues.
3. Data visualization: Develop a graphical dashboard to display AQI and atmospheric pollutant contributions and the tendency over the preceding and following 5 days.

Fig 1: Examples of internal validation plots showing time series of AQI values computed from observed and modelled values together with information on the dominant pollutant.

Fig 2: Early prototype of an AQI dashboard showing CAMS forecasts in combination with observations collected by the waqi.info portal

[Ben-EJ]

We are trying to understand the audience for the dashboard. Is this for internal usage to measure the effectiveness of forecasts and to send alerts when the forecast and measured values have diverged too far, or more for the general public?

[mparrington]

Hi Ben,

Thanks for your question and interest in the project.

The primary use of the dashboard will be internal for day-to-day monitoring of air quality alerts in multiple locations around the world, and diagnosing forecast performance in terms of the pollutants used in the air quality index calculation. However, the dashboard should not be limited to this and any proposed development which covers both internal and external applications are anticipated and will be very welcome.

I hope this helps you to understand the requirements a bit more but please let me know if anything is still unclear or if you have any further questions.

Mark

[Ben-EJ]

How do you imagine the frontend to look, are you thinking more of a web app or more in line with Jupiter notebooks. If it is a web app, are there any frameworks ECMWF would prefer to use?

[mparrington]

The main requirement for the frontend is to included the AQI and some information on the individual contributions - similar to the example above, and there are other examples for other organizations online (e.g., http://aqicn.org/map/world/). Similar challenges in recent years have developed applications more in line with Jupyter notebooks but you can feel free to propose what you think will be effective, achievable, solutions and details can be discussed at the kick-off meeting if the challenge is selected.

[geacomputing]

Hi.

We are thinking in submitting a proposal. Any available templates for the proposal? Max and min number of pages?
Would we be given access to computing facilities of ECMWF?
Should there be a download section too, or is it web-based visualization only?
Any restrictions, in general, from a technical point of view?

Thanks for the feedback
Marco

[trakasa]

Hi Marco,
many thanks for your interest in the challenge! And great that you are thinking to submit a proposal.
As for your questions, please find my answers inline.

Any available templates for the proposal?
AT: There isn't any template you need to use, as challenges differ quite a lot.
I would recommend though to check the FAQs, they provide some tips & recommendations.
E.g.

• it is well structured and precise
• the solution proposed is technically feasible within 4 months
• it contains a clear timeline with milestones and deliverables defined

The Q on the evaluation process provides also input.

Additionally you might want to listen to any of the Q&A webinar recordings on the Code for Earth Youtube channel.

Max and min number of pages?
AT: There is no min or max number of pages. The submission and proposed solution should mirror the aspects mentioned above.

Would we be given access to computing facilities of ECMWF?
AT: Yes, if necessary we can provide computing facilities through the European Weather Cloud or WEkEO.

Should there be a download section too, or is it web-based visualization only?
Any restrictions, in general, from a technical point of view?
AT: Unfortunately I cannot answer these two questions. I have to refer to the mentors of the challenge. Sorry for that.

I hope this helps you to get going!
All the best,
Athina
for the Code for Earth co-ordination team

[sebsteinig]

Hi,

We are considering using React for the frontend and saw that there is already a React and deployment template for challenge 12 (#5). Would it be possible to build the dashboard based on this template to make it easier to integrate it into the ECMWF ecosystem?

Thanks,
Seb

[tjohnson-scottlogic]

The code for the solution to this challenge can be found here: https://github.com/ECMWFCode4Earth/vAirify