tienlx93 / ds-project1-airbnb-seattle Goto Github PK

1. Installation
2. Project Motivation
3. File Descriptions
4. Results
5. Licensing, Authors, and Acknowledgements

There should be no necessary libraries to run the code here beyond the Anaconda distribution of Python. The code should run with no issues using Python versions 3.*.

This is my first project into the Data Science field to apply the Crips-DM (link) methodology.

For this project, I was interestested in using Airbnb listing data of Seattle to better understand:

• What is the interesting parts about the data, included pricing, size, how to create a good description for a listing...
• What aspect impact the pricing and popularity of a listing

• Seattle Airbnb analysis.ipynb: Main notebook file included all the works. This included 4 parts:

1. Part 1: Review & cleanup data
2. Part 2: Pricing overview
3. Part 3: How to get good impression by title, description & user reviews
4. Part 4: What aspect affect pricing & user review

• data/listings.csv: detail of the listing, include detail of rental, place, pricing, user review
• data/reviews.csv: detail of all review of places
• data/calendar.csv: detail of pricing & avaiability of listings by day
• output: folder containing output tables, images

Throughout the project file Seattle Airbnb analysis.ipynb it will help to demostrate the CRIPS-DM method.

• Business Understanding

  • Question 1: When is the busiest time to visit Seatle? Where is the common place to be visited?
  • Question 2: Based on the description & review, what we can learn to create a more attractive listing?
  • Question 3: What aspect impacted the most on nighly pricing? What aspect impacted the popularity of a listing, based on number of reviews per month?

• Data Understanding: The provided data included:

  • listings.csv: detail of the listing, include detail of rental, place, pricing, user review
  • reviews.csv: detail of all review of places
  • calendar.csv: detail of pricing & avaiability of listings by day

• Prepare Data: Through the first part of 'Review & cleanup data', we run through some clean up & prepare like:

  • Cleanup the numeric & pricing value for a clean numberic output
  • Remove the unnessary data
  • Create the statistic tables, correlation matrix, data distribution for better understanding of data
  • For the question 1, we merge the listings.csv and calendar.csv to see the pricing distribution. It is covered on the Part 2 of the project
  • For the question 2, we merge the listings.csv and reviews.csv for the review & descritions of a listing. We applied the words deviding, then count the appearance of each word, to compare the difference between the most attractive listing (based on number of reviews per month) and all listing. The findings are available on Part 3 of the project.

• Modeling

  • Fit model: After clearing & enrich data, create columns for categorical values, we fit the model into the standard linear regression model. This is done in the 4th part on the notebook.
  • Validate the model: After changing the parameter, we removed some columns that create overfitting data, only focus on most interesting data
  • After apply the modeling, we can answer Question 3, what aspect affect the most on pricing and popularity. The result is available on the Part 4 of the notebook.

• Evaluation: The main findings of the project can be found at the post available here.

The data inside data folder is credited to Airbnb for the. You can find the Licensing for the data and other descriptive information at the Kaggle link available here.

The notebook is created for educational purpose. Please freely to contribute, remix and make it your own work / with or without any credit to original works.