airbnbFeatureEngineering1.Rmd

```{r}
# Load packages

options( java.parameters = "-Xmx6g")

library(dplyr)
library(tidyr)

# Create example data frame

library(caret)

library('tidyverse')
library('tidymodels')
library('plotly')
library('skimr')
library(ggplot2)
library(mlr)
library(xgboost)
library(splitstackshape)
library(janitor)

train <- read_csv('airbnbTrain.csv')

#creating column with count of amenitites

train$amenities <- substr(as.character(train$amenities),
                       start= 2, 
                       stop= nchar(as.character(train$amenities) )-1 )

#creating column of number of days the person has been host
train$host_sinceD<-Sys.Date()-train$host_since
train$host_sinceD<-as.numeric(train$host_sinceD)


#creating dummy columns for each of the amenities.

train<-cSplit_e(train, "amenities", type = "character", fill = 0)

#creating new dataframe with just columns of amenities.
train2<-train[,-c(1:67)]

train2$host_sinceD[is.na(train2$host_sinceD)] <- mean(train2$host_sinceD, na.rm = TRUE)


# cleaning the names of columns
train2<-clean_names(train2)



#removing _number pattern from columns. This pattern is beacuse of duplicated columns created after cleaning the names. Thus 
#columns such as amenities_fixed_grab_bars_for_shower_2 is converted into amenities_fixed_grab_bars_for_shower, result of this is duplicate named column.
names(train2) <- sub("\\_\\d+{1}$", "", names(train2))

# adding columns with duplicate names and dropping the duplicatede column.

dat<-train2

indx <- sapply(dat, is.numeric)#check which columns are numeric
nm1 <- which(indx)#get the numeric index of the column
indx2 <- duplicated(names(nm1))|duplicated(names(nm1),fromLast=TRUE)
nm2 <- nm1[indx2]
indx3 <- duplicated(names(nm2))
dat[nm2[!indx3]] <- Map(function(x,y) rowSums(x[y]), list(dat),split(nm2, names(nm2)))
datN <- dat[ -nm2[indx3]]


names(datN)


```