Introducing 

Prezi AI.

Your new presentation assistant.

Refine, enhance, and tailor your content, source relevant images, and edit visuals quicker than ever before.

Loading…
Transcript

Data Analysis

on Hotel Reviews

Presented by Tal Maimon, Ben Gambash, Pamela ezra

THANK YOU!

OUR COLLEGE

OUR OFFICE

College of Managment and Academic Studies , Rishon Le'zion.

TEAM

TEAM

TEAM

Economic and Management student.

Dean's list of 2017.

"Cooking ninja"

Tal Maimon

Business student, financial and stock market speciality.

Dean's list of 2017.

"You'll never know your limit unless you push yourself to them"

Ben Gambash

Economic and Management student.

Teamleader of Economic Forum.

"Life isn't about finding yourself, is about creating yourself"

Pamela Ezra

HOTEL REVIEWS

DATA

Our data contains 515,000 costumer reviews and

scoring of 1493 hotels across Europe.

The geographical location of hotels is also provided.

columns: hotel adress, review date, average score, hotel name, reviewer nationality, negative review, review total negative word counts, positive review, review total positive word counts, reviewer score, total number of reviews reviewer has given, total number of reviews, tags, days since review, additional number of scoring, lat, lng.

PROBLEMS

PROBLEMS

01.

Duplicated values

02.

Missing data

03.

Long codes that took A LOT of time

SOLUTIONS

SOLUTIONS

We cleaned all the duplicated values

We created a dict of latitude and longtitude using

the site: https://www.latlong.net/

We used Debugger to check if there's a problem in the code before running it

ANALYSIS

ON DATA

somthing that will matter

something that will change

something we'll be proud of

WHAT WE WANTED TO FIND

somthing new

OUR HYPOTHESIS

1. There's a diffenence in the reviews' scores and the amount of reviews during the year.

2. There are nations that bring more reviews then others.

3. There's a difference in the scores between business and liesure purpose.

4. There's a difference between the reviewers score and the sentiment score of their reviews.

Small things we did durring the process:

difrrences on scores between business ans leisure visitors

Top reviewed hotels

average amount of reviews per date

reviewers nationality

correlation between variables

average score per date

Let's go deeper

we tried to predicted if a visitor came for a business of leisure purpose , our occurancy hit 81%

we created a KPI that reflects the positive ratio of each review

what it means excatly - positive and negative review?

positive words

we created a function that deletes the duplicated values in order to see more clearly what are the positive words and the negative words....

negative words

we can see that we have THE SAME WORDS in both of them....

we can see that our prediction accurancy is 86%! and when we tried use it it succeed recognize positve and negative reviews!

since there is a difference in the positive and negative reviews - we wanted to predict whether the review is positive or negative

now we can really see what matters

another prediction we used was a decision tree:

it's accurancy is 78%

Further more with sentiment analysis:

we decided to use VADER library

let's see it more clearly...

we can see the differences between the reviewers' score and the sentiment score

each review had a sentiment score scale from -1 to 1 based on vader parameters

CONCLUSIONS

דso after we did all that..

what we can do with it?

CONCLUSIONS

CONCLUSIONS

trendy positive and negative words can show us what's really matters for visitors when they come to the hotel, by using the positive words we can organize a personal trip based on what other people liked about the area. by using the negative words the hotel will know how to improve.

"

if we know what are the seasons costumers come to our hotel we can prepare correctly and organize our manpower.

we can know who our clients are, and then reach them using personal advertisment.

We have 4 different types of review analysis, we can combine them in order to create a sentiment parater that will help the hotel analyze their costumers review, in that way they'll know how to optimize themselves and reach more clients. We can also create an interactive tool that can do the analysis for the hotel and provide only the main words and their costumer's score.

"

if we have more business or leisure clients we can optimize our hotel features.

by knowing who are the costumers that reviewed the most, we know which visitors needs to be treated in a certain way so our overall score will rise.

Q

QUESTIONS?

Learn more about creating dynamic, engaging presentations with Prezi