Data based model for predicting COVID-19 morbidity and mortality in metropolis

Area of Science:

• Epidemiology
• Data Science
• Environmental Health

Background:

• The COVID-19 pandemic necessitates robust scientific analysis for public health decision-making.
• Effective strategies require understanding the interplay between environmental factors and disease transmission.
• Previous models often lacked city-specific environmental variable integration.

Purpose of the Study:

• To develop and validate a data mining methodology for predicting COVID-19 evolution in metropolises.
• To identify key air quality and meteorological variables correlated with COVID-19 cases and deaths.
• To compare the performance of different forecasting algorithms for epidemic prediction.

Main Methods:

• Application of data mining techniques to epidemiological data from São Paulo, Rio de Janeiro, and Manaus.
• Statistical analysis to determine significant environmental predictors of COVID-19.
• Cluster analysis to identify optimal input variables for forecasting models.
• Development and comparison of two distinct algorithmic forecasting models.

Main Results:

• Identified significant correlations between environmental variables and COVID-19 outcomes, varying by city.
• Discovered specific environmental indicators: low solar radiation in Manaus and drought in São Paulo predict increased mortality.
• Developed predictive models capable of forecasting new COVID-19 cases and deaths.

Conclusions:

• The data mining methodology effectively predicts COVID-19 trends and identifies critical environmental factors.
• Environmental variable relationships with COVID-19 are geographically specific.
• The approach is adaptable for other cities and epidemic diseases, offering a versatile public health tool.