Analytics Assignment 2023
Introduction
When tasked with analyzing the data-set hotel_bookings, we first had to look into each variable. Understanding the variables you’re working with is the essential first step when it comes to analyzing data.
The following variables have been taken from the hotel_bookings data-set.
year
gender
age
origin_country
destination_country
destination_city
hotel
hotel_rating
payment_method
num_people
num_days
num_rooms
booking_price
discount_pct
Having understood each of the variables, it was then clear that we needed to set out sections in order to breakdown this data. The sections in this document are;
Section One - Who’s in our data set?
Section Two - Country Analysis
Section Three - Price Analysis
Section Four - Payments
Section One - Who’s in our data set?
When looking at the data set hotel_bookings we can see several variables. Variables ranging from customer data to hotel data. Prior to deep diving into intricate graphs and tables, I believe it’s important to get a good representation of who is travelling. To do this we first look at age.
Figure 1
Above is a histogram showing the Age distribution across our data set. This shows us how many people from each age bracket have booked a hotel. We can see that there are two outlying age brackets on this histogram. With people between 55-65 years old having the least number of hotel bookings. This is an interesting find, as there are multiple ways to analyse why this may be. On initial thought one might consider that this is lower than expected as people in that age bracket are closer to retirement and would be expected to have more disposable income, being able to afford to go on more trips away. However, under further consideration it could be lower due to a lack of technological ability. Meaning that people that are elderly might find booking hotels online or over the phone a tedious task.
We can see that the other outlying age demographic is those under the 20 years old bracket. When thinking about why this may be lower than the other demographics, we consider that people in this demographic are likely to be on the lower income scale. This may be due to people in this demographic still being in college, with many having completed college recently and entering the workforce on an entry level salary.
To further investigate into this thought process, we have grouped the ages into brackets. To investigate the above we looked at what age bracket has spent the most in terms of booking_price.
Figure 2
As seen above, the 19-24 year old age bracket has spent the most in terms of the Sum of booking_price. This shows us that unexpectedly the 19-24 age bracket has spent over €700,000 more than the 25-29 year old age bracket. A bracket with a drastically larger demographic. However, our assumption regarding the older population has proven through as the oldest age demographic has the lowest amount spent in regards to bookings.
Section Two - Country Analysis
Looking at our data as a whole file we can very easily see what country’s our data set are coming from prior to travelling and what country’s that our data set are planning a trip to. However, this information is visible on a individual level. Looking at this data as a whole is important, finding out what the most popular country relationships are can be crucial to marketing executives. Country relationships being what pair of countries (Origin and Destination) has the most travelers. This information can then be used by marketers to target holiday packages to users in each origin country about the relative popular destination countries.
Below is a table showcasing the Top 15 Origin to Destination Countries. As we can see the most booked trip is from Thailand to Nepal, closely followed by Indonesia to Nepal.
Table 1
| Country of Origin | Destination Country | Count |
|---|---|---|
| Thailand | Nepal | 660 |
| Indonesia | Nepal | 655 |
| Malaysia | Iceland | 653 |
| Singapore | New Zealand | 646 |
| Singapore | Iceland | 644 |
| Malaysia | China | 643 |
| Malaysia | Ireland | 637 |
| Thailand | China | 636 |
| Indonesia | Colombia | 635 |
| Singapore | Egypt | 633 |
| Indonesia | Israel | 631 |
| Singapore | Canada | 630 |
| Thailand | Italy | 629 |
| Singapore | Colombia | 627 |
| Malaysia | New Zealand | 625 |
Although the above table is important and gives us a great deal of information about the relationship between pairs of countries, it’s also important to look into individual countries. Specifically which countries have the most amount of travelers, both origin and destination countries.
Figure 3
As we can see from the barchart “Origin Country Summary”, Thailand is the country with the largest population of travelers. However, this is closely followed by Malaysia and Singapore.
When looking at the Figure 4, a scatter-plot named “Destination Country Summary” we can see that New Zealand is the most traveled too country out of all the destination countries. If we refer back to Figure 3 and Table 1 we can see that although Thailand is the country that has traveled the most, it is people from Singapore that have visited New Zealand the most. This outcome is what we would expect from this data, as in second place within Figure 4 is Nepal. Nepal is the country that holds the two top spots of Table 1, with Thailand and Indonesia.
Figure 4
Section Three - Price Analysis
When thinking about price, we want to look at the average booking price by destination country to see which country costs the most vs which costs the least. As we can see from Table 2 below, Germany on Average has the most expensive hotels while Japan seems to have the cheapest. This is interesting as we would presume that Japan, being a cheaper option on average would have a lot of visitors compared to the other destination countries. However, if we refer back to Figure 4 we can see that Japan almost has the lowest number of hotel bookings. This low price may be due to the fact that hotels are lowering their prices in order to try boost the number of visitors to the country.
Table 2
| Destination Country | Mean Booking Price |
|---|---|
| Germany | 217.6442 |
| Israel | 217.4032 |
| Iceland | 216.9644 |
| Maldives | 216.4695 |
| Colombia | 215.4123 |
| Kenya | 214.8296 |
| Iran | 214.6624 |
| Italy | 214.4641 |
| Ireland | 214.0953 |
| New Zealand | 214.0775 |
| Canada | 213.9955 |
| Mexico | 213.7974 |
| China | 213.2439 |
| India | 212.8426 |
| Denmark | 212.3144 |
| Egypt | 212.2823 |
| Brazil | 212.2660 |
| France | 212.2615 |
| Nepal | 211.4106 |
| Japan | 210.7524 |
We can see from the data-set hotel_bookings that the price of the hotel room has some correlation with the hotel rating. However, we also know that discount percentage is another variable that changes the hotels booking price as well as a consumers decision making process when booking.
While price has a large effect on which hotel consumers book, the rating also has an incredibly large impact on the consumer. So we decided in Figure 5 to look into the average hotel rating for each country. We displayed this on a scatter-plot. As you can see the Maldives has on average the highest rated hotels with Nepal having on average the lowest rated hotels. We can also see an average line through the visualization marked in red. This result is interesting as Nepal is one of countries witht he highest number of visitors from our data-set (Figure 4). With such a high visitor count, we would’ve presumed that the the average hotel rating would’ve been drastically higher than last place.
Figure 5
This has lead us to wonder what changes the discount percentage offered to consumers on each booking. Our main presumption was that the more people being booked into the booking, the higher the discount would be. To investigate this presumption we looked at Figure 6, a boxplot showcasing discount percentage by the number of people staying.
Surprisingly our presumption was wrong. We found that the average, min and max discount percentage was the very close depending on how many people were staying. This would be a possible suggestion to marketing teams, offering discounts based on number of people staying might be more beneficial in terms of separating you from the competition in terms of bookings. What you may loose financially in the booking, you would expect to make up for when it comes to people drinking and dining in the hotel facilities.
Figure 6
To further look into the booking price of the hotels. We thought what else could have an effect on the price. Naturally we presumed that the higher the rating of the hotel, the higher the booking price would be. To investigate this we created a scatter-plot in Figure 7, this scatter-plot looks at the relationship between the average rating each hotel has vs the average booking price for the respective hotel.
Figure 7
We can see from Figure 7 that our presumption about their being a relationship between hotel rating and booking price was right. As expected there is a positive linear association with our two variables. Although there will always be outlyers in the data, we can clearly see that for the majority of the data the higher the hotel rating the higher the average booking price.
Section Four - Payments
We can see from the data-set hotel_bookings that there is a large variety of payment methods used for bookings. Our curiosity came from what payment method is most used by year. Our presumption was that over the years internet banking and debit would drastically over take every other option. To investigate this we created a line graph in Figure 6, showcasing the performance of each payment method for each year.
Figure 8
We can see in Figure 8 that as we expected, internet banking really took off in 2011 drastically out performing the other payment methods. However, debit card very quickly caught up with internet banking with the digital options completely out performing wallet the further along the timeline we look. This is expected as society has been drastically moving cashless .
Meme
