med_business_clusters.R

library(data.table); library(dplyr); library(tidyr); library(ggplot2);library(NbClust)

## Read Data
data <- read.table("~/Downloads/business_users.csv",sep=",",header=T)
names(data) <- c("user_id","team_members","profiles","months_paying","plan_days","updates","updates_per_day_per_profile")

## Prep Data
data[is.na(data)] <- 0

## Separate Variables We're Interested In
features <- data %>%
  select(-user_id,-plan_days,-updates,-months_paying)

## Find how many clusters to use with wssplot
wssplot <- function(data, nc=15, seed=1234){
  wss <- (nrow(data)-1)*sum(apply(data,2,var))
  for (i in 2:nc){
    set.seed(seed)
    wss[i] <- sum(kmeans(data, centers=i)$withinss)}
  plot(1:nc, wss, type="b", xlab="Number of Clusters",
       ylab="Within groups sum of squares")}

wssplot(features)

## NbClust
nc <- NbClust(features, min.nc=2, max.nc=15, method="kmeans")
table(nc$Best.n[1,])

barplot(table(nc$Best.n[1,]),
        xlab="Numer of Clusters", ylab="Number of Criteria",
        main="Number of Clusters Chosen")


## K-means Clustering
set.seed(1234)
results <- kmeans(features,3)
results


## Add cluster to original dataset
data$cluster <- as.factor(results$cluster)


## Plot number of profiles and team members for each cluster
ggplot(data,aes(x=profiles,y=team_members,color=cluster)) +
  geom_point(size=) + 
  scale_x_continuous(limits=c(0,50)) +
  labs(x="Number of Profiles",y="Number of Team Members") + 
  theme_minimal()