Introduction
This presentation is one half of the assignemnt of week 4, Developing Data Products course from Coursera. In this presentation we will pitch for the shiny application we have developped, which can be found here.
In this shiny application we are going to predict the Miles/Gallon from the mtcars dataset taking the values from other features in the dataset interactively.
Analysis
In the shiny app we have considered the mtcars dataset for our analysis.
The variable “am” in mtcars represents the transmission mode of the vehicle i.e “1” for Manual and “0” for automatic transmission.
Let’s convert the “am” variable to a factor variable replacing the binary 1 & 0 with “Auto” and “Manual” respectively.
mtcars$am <- factor(ifelse(mtcars$am == 1, "Manual", "Auto"))
We have shown the relationship of all the features against Miles/Gallon
- Box Plots for the features am, vs, gear, carb, cyl
- and scatter plots for the features disp, hp, drat, wt, qsec
Regression
In the app we have shown linear regression between the variables . A sample code is given below.
model <- lm(formula = mpg ~ cyl + hp + wt + am, data = mtcars)
with(mtcars, {
plot(model)
abline(model, col=2)
}
)
## Warning in abline(model, col = 2): only using the first two of 5 regression
## coefficients
Forecast
We have trained model now we predict the Miles/Gallon according to inputs. In shiny app this has been done interactively.
An example shown below.
## Loading required package: lattice
## Loading required package: ggplot2
customTrainControl <- trainControl(method = "cv", number = 10)
trModel <- train(mpg ~ cyl + hp + wt + am, data = mtcars, method = "lm", trControl = customTrainControl)
predict(trModel, newdata = mtcars[1:10,])
## Mazda RX4 Mazda RX4 Wag Datsun 710 Hornet 4 Drive
## 23.58005 22.91539 26.27647 20.55114
## Hornet Sportabout Valiant Duster 360 Merc 240D
## 16.85255 20.03731 14.76713 23.30427
## Merc 230 Merc 280
## 22.58514 19.64032