3D Plotting is Cool!

This plot aims to locate some key features that might predict the success or failure of a kickstarter project. The data for this analysis comes from Cathie So’s Kickstarter Project Statistics. We’ll only be analysing the 4000 most backed kickstarter projects.

We’ve made some modifications to the core data. And we’ve created 3 classes of projects: Games, Movies and Technology. This will help determine if the category of project has an imact on performance.

Technically, this classification color coding will actually make this a 4D plot.

Formatting the Data

library(stringr)

# Data files must be downloaded into your working directory
x <- read.csv("most_backed.csv")
caliVar <- grepl(" CA$",x$location)
x <- cbind(x, caliVar)

# Convert all pricing to USD

x[grep("^aud$", x$currency),]$goal <- (x[grep("^aud$", x$currency),]$goal * 0.73)
x[grep("^gbp$", x$currency),]$goal <- (x[grep("^gbp$", x$currency),]$goal * 1.23)
x[grep("^cad$", x$currency),]$goal <- (x[grep("^cad$", x$currency),]$goal * 0.75)
x[grep("^eur$", x$currency),]$goal <- (x[grep("^eur$", x$currency),]$goal * 1.04)
x[grep("^sek$", x$currency),]$goal <- (x[grep("^sek$", x$currency),]$goal * 0.11)
x[grep("^nzd$", x$currency),]$goal <- (x[grep("^nzd$", x$currency),]$goal * 0.69)
x[grep("^dkk$", x$currency),]$goal <- (x[grep("^dkk$", x$currency),]$goal * 0.14)
x[grep("^chf$", x$currency),]$goal <- (x[grep("^chf$", x$currency),]$goal * 0.97)

x[grep("^aud$", x$currency),]$amt.pledged <- (x[grep("^aud$", x$currency),]$amt.pledged * 0.73)
x[grep("^gbp$", x$currency),]$amt.pledged <- (x[grep("^gbp$", x$currency),]$amt.pledged * 1.23)
x[grep("^cad$", x$currency),]$amt.pledged <- (x[grep("^cad$", x$currency),]$amt.pledged * 0.75)
x[grep("^eur$", x$currency),]$amt.pledged <- (x[grep("^eur$", x$currency),]$amt.pledged * 1.04)
x[grep("^sek$", x$currency),]$amt.pledged <- (x[grep("^sek$", x$currency),]$amt.pledged * 0.11)
x[grep("^nzd$", x$currency),]$amt.pledged <- (x[grep("^nzd$", x$currency),]$amt.pledged * 0.69)
x[grep("^dkk$", x$currency),]$amt.pledged <- (x[grep("^dkk$", x$currency),]$amt.pledged * 0.14)
x[grep("^chf$", x$currency),]$amt.pledged <- (x[grep("^chf$", x$currency),]$amt.pledged * 0.97)

x$currency <- "usd"

# Calculate important ratios

x$pledge.ratio <-x$amt.pledged/x$goal
x$per.backer.pledge <-x$amt.pledged/x$num.backers

# Create 3 classes of project and combine into new dataset

xmovie <- x[grepl("movie", x$blurb),]
xmovie$kickType <- "movie"
xgame <- x[grepl("game", x$blurb),]
xgame$kickType <- "game"
xtechnology <- x[grepl("technology", x$blurb),]
xtechnology$kickType <- "technology"

kicktypex <- rbind(xmovie, xgame, xtechnology)
kicktypex$kickType <- as.factor(kicktypex$kickType)

# Create new factors for further analysis

kicktypex$num.tiers <- str_count(kicktypex$pledge.tier, ',') + 1
kicktypex$per.backer.pledge.mse <- (kicktypex$per.backer.pledge - mean(kicktypex$per.backer.pledge))^2

Now The Magic Happens

You can also embed plots, for example:

library(plotly)
## Loading required package: ggplot2
## 
## Attaching package: 'plotly'
## The following object is masked from 'package:ggplot2':
## 
##     last_plot
## The following object is masked from 'package:stats':
## 
##     filter
## The following object is masked from 'package:graphics':
## 
##     layout
library(webshot)
plot_ly(x = ~kicktypex$num.tiers, y = ~kicktypex$goal, z = ~kicktypex$per.backer.pledge, type = "scatter3d", color = ~kicktypex$kickType)
## No scatter3d mode specifed:
##   Setting the mode to markers
##   Read more about this attribute -> https://plot.ly/r/reference/#scatter-mode
## Warning in arrange_impl(.data, dots): '.Random.seed' is not an integer
## vector but of type 'NULL', so ignored

That… was… awesome.

awesome

awesome