hotel_booking_demand_wk3.Rmd

---
title: "Hotel Booking Pattern & Analysis"
author: "Prasham Bhuta"
date: "30/03/2020"
output:
  pdf_document: default
  html_document: default
---

```{r setup, include=FALSE}
knitr::opts_chunk$set(echo = TRUE)
```

## Analysis of data received from Room Booking Platform

Online platforms such as trivago, goibibo, makemytrip etc. are used for booking hotel rooms, here is a dataset of rooms booked from such platform.

### Tasks

* To understand the pattern of bookings, and the general trends followed by users.
* Create a report, as being a part of the platform, for the marketing team.

## Let's get started

### Importing necessary libraries
```{r, echo=TRUE, message=FALSE}
library(tidyverse)
```

### Import data from the csv
```{r}
datas <- read.csv("dataset/hotel_bookings.csv")
```

### Get the str of the data
```{r}
str(datas)
```

### Understanding the columns

The data looks clean enough, with proper column headers, as well

* Hotel has two types:
    * Resort hotel
    * City hotel
    
* Is_cancelled:
    * "1" if the booking is cancelled

* lead_time
    * No of days between booking and booked date

* Arrival : year, month, week_number, day

* Stay
    * No of weekend nights
    * No of week nights
(because of price difference during the weekends)

* No of people: adults, children, babies

* Meal booked has 5 types
    * BB - Bed & Breakfast
    * FB - Full Board (Breakfast, Lunch & Dinner)
    * HB - Half Board (Breakfast + 1 other (dinner or lunch, mostly dinner))
    * SC - No Meal package
    * Undefined - No Meal package

* Country (self - explanatory)

* market_segment (group of people who share common characteristic)

* distribution_channel (intermediaries between users and hotel booking eg. websites, travel agents, tour operators)

* is_repeated_guest (has previous booking)

* previous_cancellations (has previously cancelled a booking)

* reserved_room_type (type of room reserved)

* assigned_room_type (type of room assigned, due to high volume this can differ from reserved_room_type)

* booking_changes (no of times changes have been made to the booking)

* deposit_type
    * No Deposit 
    * Non Refund - deposit of value equals total cost
    * Refundable - value under the total cost of stay
    
* agent
    * ID of travel agency that made the booking
    
* Company
    * ID of the company responsible for booking or payment
    
* days_in_waiting_list (no of days before the booking was confirmed)

* customer_type
    * Contract
    * Group
    * Transient
    * Transient-party
    
* adr (average daily rate)
    * adr = (sum_of_all_expenses)/(total_nights_of_stay)
    
* required_car_parking_spaces

* total_of_special_requests

* reservation_status
    * Canceled
    * Check-Out
    * No Show - Customer did not show up

* reservation_status_date
    * Date when the final changes to the entry was made.
    
### Calculating the NA, values
```{r}
colSums(is.na(datas))
```
There are no such columns with na values that need to be removed or edited

### Hotel

There are just two types of hotel, Resort & City, so a basic barplot would give the idea of the percentage of booking.

```{r}
counts <- prop.table(table(datas$hotel))
barplot(counts, col = c('lightblue','lightgreen'), main = "Type of Hotel")
```

#### Analysis

City hotels are twiced as much booked compared to Resort hotels, following reasons can be derived for that.

* City hotels are better options for corporate bookings, and business purposes

* Resort hotels can be a good option or larger parties.

### Cancelled bookings

To understand what percentage of bookings are cancelled.
```{r}
cancelled <- prop.table(table(datas$is_canceled))
barplot(cancelled, main = "Percentage of bookings cancelled", names.arg = c("Not Cancelled", "Cancelled"), col = c("lightblue", "red"))
```