文件列表:

EDA(Group)_Hotel_Booking_Analysis_Capstone_Project.ipynb (637037, 2022-05-12)
Final_EDA(Group)_Hotel_Booking_Analysis_Capstone_Project.ipynb (932137, 2022-05-12)
eda powerpoint.pdf (1071019, 2022-05-12)

# Hotel Bookings Exploratory Data Analysis ## Objective We are provided with a hotel bookings dataset. Out main objective is perform EDA on the given dataset and draw useful conclusions about general trends in hotel bookings and how factors governing hotel bookings interact with each other. ## Dataset We are given a hotel bookings dataset. This dataset contains booking information for a city hotel and a resort hotel. It contains the following features. ``` - hotel: Name of hotel ( City or Resort) - is_canceled: Whether the booking is canceled or not (0 for no canceled and 1 for canceled) - lead_time: time (in days) between booking transaction and actual arrival. - arrival_date_year: Year of arrival - arrival_date_month: month of arrival - arrival_date_week_number: week number of arrival date. - arrival_date_day_of_month: Day of month of arrival date - stays_in_weekend_nights: No. of weekend nights spent in a hotel - stays_in_week_nights: No. of weeknights spent in a hotel - adults: No. of adults in single booking record. - children: No. of children in single booking record. - babies: No. of babies in single booking record. - meal: Type of meal chosen - country: Country of origin of customers (as mentioned by them) - market_segment: What segment via booking was made and for what purpose. - distribution_channel: Via which medium booking was made. - is_repeated_guest: Whether the customer has made any booking before(0 for No and 1 for Yes) - previous_cancellations: No. of previous canceled bookings. - previous_bookings_not_canceled: No. of previous non-canceled bookings. - reserved_room_type: Room type reserved by a customer. - assigned_room_type: Room type assigned to the customer. - booking_changes: No. of booking changes done by customers - deposit_type: Type of deposit at the time of making a booking (No deposit/ Refundable/ No refund) - agent: Id of agent for booking - company: Id of the company making a booking - days_in_waiting_list: No. of days on waiting list. - customer_type: Type of customer(Transient, Group, etc.) - adr: Average Daily rate. - required_car_parking_spaces: No. of car parking asked in booking - total_of_special_requests: total no. of special request. - reservation_status: Whether a customer has checked out or canceled,or not showed - reservation_status_date: Date of making reservation status. ``` - Total number of rows in data: 119390 - Total number of columns: 32 ## Data Cleaning and Feature Engineering ### (1) Removing Duplicate rows All duplicate rows were dropped. ### (2) Handling null values - Null values in columns `company` and `agent` were replaced by 0. - Null values in column `country` were replaced by 'others'. - Null values in column `children` were replaced by the mean of the column. ### (3) Converting columns to appropriate data types - Changed data type of `children`, `company`, `agent` to int type. - Changed data type of `reservation_status_date` to date type. ### (4) Removing outliers - One outlier was found in the `adr` column. Simply dropped it. ### (5) Creating new columns - Created new column `total_stay` by adding `stays_in_weekend_nights`+`stays_in_week_nights`. - Created new column `total_people` by adding `adults`+`children`+`babies`. ## Exploratory Data Analysis Performed EDA and tried answering the following questions: ``` Q1) Which agent makes the most no. of bookings? Q2) Which room type is in most demand and which room type generatesthe highest adr? Q3) Which meal type isthe most preffered meal of customers? Q4) What isthe percentage of bookings in each hotel? Q5) Which is the most common channel for booking hotels? Q6) Which are the most busy months? Q7) From which country most of the guests are comin ? Q8) How long do people stay at the hotels? Q9) Which hotel seems to make more revenue? Q10) Which hotel hasa higher lead time? Q11) What is preferred stay length in each hotel? Q12) Which hotel has higher bookings cancellation rate. Q13) Which hotel hasa high chance that its customer will return for another stay? Q14) Which channel is mostly used forthe early booking of hotels? Q15) Which channel hasa longer average waiting time? Q16) Which distribution channel brings betterrevenue-generatingg deals for hotels? Q17) Which significant distribution channel has the highest cancellation percentage? Q18) Doesa longer waiting period or longer lead time causes the cancellation of bookings? Q19) Whether not getting allotted the same room type as demand is the main cause of cancellation for bookings? Q20) Does not alloting the same room as demanded affect adr? Q21) What is the trend of bookings within a month? Q22) Which types of customers mostly make bookings? ``` Mainly performed using Matplotlib and Seaborn library and the following graph and plots had been used: - Bar Plot. - Histogram. - Scatter Plot. - Pie Chart. - Line Plot. - Heatmap. - Box Plot ### Univariate Analysis: Performed univariate analysis and made following conclusions: ``` 1.) Agent no. 9 has made most no. of bookings. 2.) Most demanded room type is A, but better adr generating rooms H, G and C. Hotels should increase the no. of room types A and H to maximise revenue. 3.) Most popular meal type is BB(Bed and Breakfast). 4.) Around 60% bookings are for City hotel and 40% bookings are for Resort hotel, therefore City Hotel is busier than Resort hotel. 5.) Guests use different channels for making bookings out of which most preferred way is TA/TO. 6.) July- August are the most busier and profitable months for both of hotels. 7.) Most of the guests came from european countries, with highest number of guests from Portugal. 8.) Most common stay length is less than 4 days and generally people prefer City hotel for short stay, but for long stays, Resort Hotel is preferred. ``` ### Bivariate Analysis : We tried to answer following questions ``` 1.) Overall adr of City hotel is slightly higher than Resort hotel and no. of bookings of City hotel is also higher than Resort hotel. Hence, City hotel is makes more revenue. 2.) City hotel has slightly higher median lead time. Also median lead time is significantly higher for both hotels, this means customers generally plan their hotel visits way early. 3.) Almost 30 % of City Hotel bookings got canceled. 4.) Both hotels have very small percentage that customer will repeat, but Resort hotel has slightly higher repeat % than City Hotel. 5.) TA/TO is mostly used for planning Hotel visits well ahead of time. 6.) While booking via TA/TO one may have to wait a little longer to confirm booking of rooms. 7.) GDS channel brings higher revenue generating deals for City hotel, in contrast to that most bookings come via TA/TO. City Hotel can work to increase outreach on GDS channels to get more higher revenue generating deals. 8.) TA/TO has highest booking cancellation %. Therefore, a booking via TA/TO is 30% likely to get cancelled. 9.) Longer lead time has no affect on cancellation of bookings. 10.) Not getting same room as demanded is not the case of cancellation of rooms. A significant percentage of bookings are not cancelled even after getting different room as demanded. 11.) Not getting same room do affects the adr, people who didn't got same room have paid a little lower adr. 12.) Arrivals in hotels increases at weekends and also the avg adr tends to go up as month ends. 13.)Moslty bookings are done by couples(bookings have two adults.) ``` ## Conclusion ``` (1) Around 60% bookings are for City hotel and 40% bookings are for Resort hotel, therefore City Hotel is busier than Resort hotel. Also the overall adr of City hotel is slightly higher than Resort hotel. (2) Mostly guests stay for less than 5 days in hotel and for longer stays Resort hotel is preferred. (3) Both hotels have significantly higher booking cancellation rates and very few guests less than 3 % return for another booking in City hotel. 5% guests return for stay in Resort hotel. (4) Most of the guests came from european countries, with most of guests coming from Portugal. (5) Guests use different channels for making bookings out of which most preferred way is TA/TO. (6) For hotels higher adr deals come via GDS channel, so hotels should increase their popularity on this channel. (7) Almost 30% of bookings via TA/TO are cancelled. (8) Not getting same room as reserved, longer lead time and waiting time do not affect cancellation of bookings. Although different room allotment do lowers the adr. (9) July- August are the most busier and profitable months for both of hotels. (10) Within a month, adr gradually increases as month ends, with small sudden rise on weekends. (11) Couples are the most common guests for hotels, hence hotels can plan services according to couples needs to increase revenue. (12) More number of people in guests results in more number of special requests. (13) Bookings made via complementary market segment and adults have on average high no. of special request. (14) For customers, generally the longer stays (more than 15 days) can result in better deals in terms of low adr. And many more conclusions. ``` ## Challenges ``` (1) There was a lot of duplicate data. (2) Data was present in wrong datatype format. (3) Choosing appropriate visualization techniques to use was difficult. (4) A lot of null values were there in the dataset.

近期下载者：

相关文件：

评论：[我要评论] [举报此文件]

收藏者：