Air Quality Prediction and Management in Poland

Challenge:

Poland faces significant air quality issues, particularly in urban areas, where pollution levels often exceed safe limits. The country, heavily reliant on coal for energy, experiences particularly poor air quality during winter months due to the widespread use of coal for heating. Broad policy measures have been implemented, but they often lack the detailed insights needed to address the root causes effectively. Effective mitigation requires a data-driven approach to accurately monitor and predict air quality, allowing for timely interventions to protect public health. Traditional methods of air quality monitoring can be slow and reactive, leading to delays in addressing pollution-related health risks.

Approach:

The project employs advanced machine learning techniques to analyse a vast array of data, including historical air quality readings, weather conditions, industrial activities, and traffic data. By training models on this data, the project aims to predict future air quality levels with greater accuracy and detail. The approach involves selecting and tuning algorithms that can process complex, multi-dimensional data and generate actionable predictions on a daily or even hourly basis. The project involves a meticulous process of translating and cleaning air quality data from Polish to English. This is followed by merging separate air quality files from each year into one comprehensive dataset. Additionally, weather data is consolidated into a single file and then joined with the air quality data to enrich the dataset. Once the data is prepared, advanced machine learning models are trained and tested to predict future levels of the Common Air Quality Index (CAQI). This supervised learning approach aims to provide precise forecasts that help mitigate pollution-related health issues and support proactive measures in Polish cities.

Outcomes:

The key outcome of this initiative is an improved ability to forecast air quality, enabling cities and communities in Poland to take proactive measures to mitigate pollution. This could include issuing health advisories, adjusting traffic flows, or implementing temporary restrictions on industrial activities. The project’s results are expected to enhance public health responses, reduce exposure to harmful pollutants, and support long-term environmental planning.

Data Sources:

In this analysis, historical air quality data from monitoring stations is examined alongside meteorological data, including temperature, humidity, and wind speed. Traffic data reflecting vehicle flow and congestion levels is also scrutinised, as well as industrial emissions reports detailing pollutant output from factories. Additionally, geographical and land use data are considered to understand the environmental context, and socio-economic data that may influence pollution levels is included.

Additional Information:

• https://www.omdena.com/blog/predicting-air-quality-in-poland-using-machine-learning
• https://datasets.omdena.com/

Point of Contact:

• https://www.omdena.com/contact-us