Art of Stat: Regression 1.5.0

The Linear Regression app provides various tools for data analysis. It allows users to generate scatterplots, establish simple linear, logistic, or exponential regression models, and conduct parameter inference including standard errors, confidence intervals, and P-values.

Additionally, the app calculates confidence intervals for mean response and prediction intervals for future responses. The graphical representation of the fitted model and intervals overlaid on the scatterplot offers a convenient visualization. Users can also access raw and standardized residuals, aiding in further analysis.

An interesting feature is the ability to color data points on the scatterplot based on a third quantitative or categorical variable, helping to uncover hidden patterns within the data.

The latest functionality includes an enhanced data entry feature through the Data Editor app, enabling users to input their own dataset, import a CSV file, or select from a variety of pre-loaded example datasets.

Moreover, users can now fit a multiple linear regression model with both quantitative and categorical predictors. The app offers:

• Scatterplot Matrix creation to explore pairwise relationships
• Clear display of the regression equation, including dummy variables
• Presentation of a regression coefficients table
• Access to summary statistics such as adjusted R-squared
• Provision of fitted values and (standardized) residuals available for download
• Capability to make predictions based on user-defined explanatory variables
• Generation of residual plots for assumption checking and outlier identification