assignment shiny intercative web page

Mark Rayfield
2nd October 2020

Dataset

Interactive shiny app based on diamonds dataset supplied with ggplot2 package.

  • 53,940 observatons across 10 variables
  • predictor variable is logically price in USD dollars
  • linar regression model is fitted to the full data set
  • Indicates carat (weight) as most importance single variable (absolute t score).
  • Other factor variables (cut, color and clarity) are non linear

Linear Regression Results

model_all <- lm(price~.,data=diamonds) 
head(summary(model_all)$coeff,10)

               Estimate Std. Error     t value      Pr(>|t|)
(Intercept)  5753.76186  396.62982   14.506629  1.352581e-47
carat       11256.97831   48.62751  231.494035  0.000000e+00
cut.L         584.45728   22.47815   26.001129 3.958257e-148
cut.Q        -301.90816   17.99392  -16.778344  5.082665e-63
cut.C         148.03470   15.48333    9.560910  1.214344e-21
cut^4         -20.79389   12.37651   -1.680110  9.294175e-02
color.L     -1952.16001   17.34177 -112.569842  0.000000e+00
color.Q      -672.05362   15.77700  -42.597060  0.000000e+00
color.C      -165.28293   14.72493  -11.224702  3.323306e-29
color^4        38.19519   13.52654    2.823722  4.748693e-03

model_red <- lm(price~carat, data=diamonds) # based only on single variable
summary(model_red)$r.squared

[1] 0.8493305

Simplified Linear model and Plot for shiny app

newC <- 1.5 # test using 1.5 carat input
pricePredicted <- predict(model_red, newdata = data.frame(carat = newC))
paste("weight in carats: ", newC, "USD price: $", round(pricePredicted,2))

[1] "weight in carats:  1.5 USD price: $ 9378.28"

plot of chunk plot

Shiny app