COVID19 Data Fitting with Linear and Nonlinear Regression

Linear, exponential, logistic, Gompertz, Gauss, Fourier models fitted to epidemiological data from the COVID-19 outbreak.

A collection of tools for fitting several general-purpose linear and nonlinear models for COVID-19 epidemiological data. The longitudinal data is obtained from the John Hopkins database (source: https://github.com/CSSEGISandData/COVID-19) and consists of: number of active cases, number of confirmed, number of fatalities, number of recovered cases. The analysis is possible for any particular country listed in the database, or for the world data as a whole. The models implemented include linear, exponential, logistic, Gompertz, fifth-degree polynomial, Gaussians and Fourier functions. The three models of the Bertalanffy class (exponential, proper logistic and Gompertz) afford a reasonable balance between reduced model complexity and goodness of fit. We implement data/model visualization in linear and logarithmic scales, for easy model comparisons.