--- 

## App Summary - Gas Milage predictor

- The App is hosted at: https://fschoen.shinyapps.io/LM_MilesPerGallon/
- It was written in R
- Toolchain: Rstudio + Shiny + ggplot2

This app predicts the average gas use per mile based on a linear regression model considering the total weight, number of cylinders and transmission type of a car. 

```{r, warning=FALSE}
library(shiny)
library(ggplot2)
```

      

        ---  

## Data

- R standard data set "mtcars"
- contains the specifications of 32 different cars
- Source: Henderson and Velleman (1981), Building multiple regression models interactively. Biometrics, 37, 391–411

```{r}
data(mtcars)
```

      

        --- 
## Model
The linear model is based on the amount of cylinders, the total weight and the transmission type. 
```{r}
# Example data from app input
input <- data.frame(slider_cyl=4, numeric_wt=2.5,checkbox_am=TRUE)

#prediction <- reactive({
prediction <-  function(){  
    m<- lm(mpg ~ cyl + wt + as.logical(am) , data=mtcars) 
    prediction_parameters<-data.frame(cyl=input$slider_cyl, 
                        wt=input$numeric_wt,am=input$checkbox_am)
    y<- predict(m,prediction_parameters)    
    #print(y)
    return(y)   
  }

prediction<-mean(prediction())
prediction
    
```
      

        --- 



## Plot
The code for creating the plot is shown below.

```{r}
plotData<- mtcars[which(mtcars$am==input$checkbox_am),]
    
    plotData$line_name<-"Original data"
    predicted_point <- data.frame(mpg = mean(prediction), cyl = input$slider_cyl,
      disp = 1,  hp = 1, drat = 1, wt = input$numeric_wt, qsec = 1,
      vs = 1, am = 1, gear = input$checkbox_am, carb = 1, line_name = 'Prediction'
    )
    plotData<-rbind(plotData,predicted_point)
    plotData$cyl<-as.factor(plotData$cyl)
    p<-ggplot(data=plotData, aes(x=wt, y=mpg, color=cyl, shape=line_name)) + 
      ggtitle("Predicted milage based on total weight, amount of cylinders \n
              and transmission type") + 
      xlab("Total Weight (in tons)") + ylab("Miles per Gallon") +
      geom_point(size=5, alpha=0.5) +
      scale_color_discrete(name="Amount of cylinders") +
      scale_shape_discrete(name="Data points")
    
```
      

        --- 
## Example Plot
This is an example plot of the app.

```{r}
p

```