Team Members

Wizair is an innovative and collaborative solution that predicts air quality based on a network of sensors. It will enable the user to avoid dangerous exposure to low quality air thanks to the live data uploaded to our mobile app and our website.

The aim is to use a group of sensors (gas, humidity, particle…) assembled in a portable box (like a tiny pollution measurement center). These sensors provide datasets over time from diverse geographical locations, which are used to predict the air quality with Machine Learning algorithms.

We chose to use a Raspberry Pi because this cheap computer could easily handle our set of sensors and could upload data in the cloud through its WiFi connectivity thanks to the API we developed. The Raspberry Pi runs a Python program which makes development very simple. The API is developed in Node JS and runs on a Linux server. Our data is stored in a Mongo DB database which is perfect for huge amount of measures.

We also began to develop a mobile app (Android and iOS platforms) using the Flutter Framework in order to avoid developing different versions of the app for each platform.

The web platform to see the measures has been developed using the Angular Framework, a performant solution to easily communicate with the API and to get data in a web browser.

What makes it innovative ?

There are numerous projects which aim to predict an air quality index, because air pollution is a major issue in big cities. On one hand, Wizair is an innovative project because it uses the latest prediction technologies, Recurrent Neural Networks. On the other hand, Wizair predicts air quality with gas measurements (nitrogen dioxide, carbon monoxide…) and meteorological measurements (humidity, temperature, wind speed…) which is more accurate than what is made today.

Wizair “boxes” are the compilation of several well chosen sensors, assembled with simple electrical links, a Raspberry Pi (affordable monocard computer). The electronics can be easily industrialized and the box too. Information (measures, localisation…) are transmitted via WiFi to the monitoring interface. Then, the machine learning comes in to deliver the prediction.

Wizair is a collaborative solution, which means that the more people use it, the more it is efficient and accurate. Because solutions to limit the environmental crisis have to be carried by everyone, Wizair is a simple way to insure quality data measures and prediction. By developing the Wizair network (installing a measure station at home or in a garden), people will extend the measurement area, but also permit the best prediction of air quality in their very local environnement. It can not only predict, but also inform on the pollution level near the station, and explain the risks and the behavior to adopt.



We are engineering students living near Paris, on a campus. Fascinated by technology, we wanted to develop a solution which could help us to understand and measure air pollution, and to be aware of the risks. We are studying electrical engineering and computer sciences, so it became obvious to learn machine learning skills, and to use our basic knowledge about electrical engineering, to contribute to the development of tomorrow’s smart cities.


Voting is closed!