In this application I developed the linear model and made the prediction for children height based on 1) averaged parent height (midparent Height), 2) child gender and 3) number of the child in the family.

Model is based on well-known Galton dataset - https://rdrr.io/cran/HistData/man/GaltonFamilies.html.

Application is available on https://pavelppz.shinyapps.io/Shiny_Galton/

Application files are available on https://github.com/PavelPPZ/Shiny-Application-and-Reproducible-Pitch/tree/main/Shiny_Galton

In the left panel you can enter 3 parameters: 1) the averaged parent height (midparent Height) by slider, 2) child gender by tick box and 3) the number of the child in the family by slider.

In the right panel you will see the results of the height prediction of the child based on linear model.

Results are presented in the text format and in the graphical form.

Results are interactive, as soon as you change the input parameters, you see the results of the child height prediction immediately.

I’m using the simple linear model with 3 parameters

        library(shiny)
        suppressMessages(library(dplyr)) 
        library(ggplot2)
        library(HistData)

data(GaltonFamilies)

LM <- lm(childHeight ~ midparentHeight+gender+childNum,GaltonFamilies)

Model is based on GaltonFamilies data.

As expected the quality of such simple model with 3 parameters is not very high. It can be assessed using model_performance function.

suppressWarnings(library(performance))
model_performance(LM)

## # Indices of model performance
## 
## AIC      |     AICc |      BIC |    R2 | R2 (adj.) |  RMSE | Sigma
## ------------------------------------------------------------------
## 3973.928 | 3973.992 | 3998.125 | 0.681 |     0.680 | 2.020 | 2.024

As you see the R-squared is low, so the quality of the model is not high.