Comparison of Nitrogen Dioxide Predictions During a Pandemic and Non-pandemic Scenario in the City of Madrid using a Convolutional LSTM Network

Abstract

Traditionally, machine learning technologies with the methods and capabilities available, combined with a geospatial dimension, can perform predictive analyzes of air quality with greater accuracy. However, air pollution is influenced by many external factors, one of which has recently been caused by the restrictions applied to curb the relentless advance of COVID-19. These sudden changes in air quality levels can negatively influence current forecasting models. This work compares air pollution forecasts during a pandemic and non-pandemic period under the same conditions. The ConvLSTM algorithm was applied to predict the concentration of nitrogen dioxide using data from the air quality and meteorological stations in Madrid. The proposed model was applied for two scenarios: pandemic (January–June 2020) and non-pandemic (January–June 2019), each with sub-scenarios based on time granularity (1-h, 12-h, 24-h and 48-h) and combination of features. The Root Mean Square Error was taken as the estimation metric, and the results showed that the proposed method outperformed a reference model, and the feature selection technique significantly improved the overall accuracy.

Reconstructing secondary data based on air quality, meteorological and traffic data considering spatiotemporal components

“…The reconstructed dataset was already used in several publications [ 2 , 5 , 6 , 7 ]. This work [2] used the reconstructed data to predict NO 2 in the next 6 h using previous 6 h data by implementing Bidirectional Convolutional Long Short-Term Memory.…”

“…This work [2] used the reconstructed data to predict NO 2 in the next 6 h using previous 6 h data by implementing Bidirectional Convolutional Long Short-Term Memory. The following work [5] used the part of reconstructed dataset, including only air quality and meteorological data, to predict NO 2 in next 1, 12, 24 and 48 h based on the previous 24 h data by implementing Convolutional Long Short-Term Memory. An additional objective was to compare the performance of predictive analysis between the pandemic (January-June 2020) and non-pandemic (January-June 2019) periods to find out if restrictions applied to curb the progress of COVID-19 impacted predicted output.…”

Graph Neural Network for Air Quality Prediction: A Case Study in Madrid

“…The main objective of the current study is to conduct spatiotemporal prediction of nitrogen dioxide (NO 2 ) in the city of Madrid using air quality, meteorological and traffic data. Some of our previous studies were also devoted to solving the above tasks [8], [9]. However, the approaches implemented there require the input to have a Euclidean or grid-like structure.…”