Group 09

Repository for Kristina Wright and Daniel Hadley group project (up to milestone 03) for STAT 547M For Milestone 04 thru Milestone 06, this repository has Daniel Hadley's project.

1. About This Repository :information_source:

This repository houses Group 09's project for STAT 547M taken in Term 2 of the 2019-2020 academic here.

Our project uses an Airbnb dataset to try and find significant factors to explain the listing prices (per night) in Barcelona, Spain.

The final report is created by meeting milestones which are linked below.

2. Navigating the Repository :file_folder:

As milestones are met, files are placed into the appropriate subfolders.

1. The data folder contains all datasets used throughout the project.
2. The docs folder contains all *.Rmd files used to create reports.
3. The images folder saves all images produced for the group project.
4. The scripts folder saves all R scripts (*.r) that are called when rendering the project.
5. The tests folder contains all tests carried out when producing the analysis.

3. Usage :computer:

1. Clone this repo.

2. Ensure the following R packages are installed:

   • tidyverse
   • here
   • docopt
   • knitr
   • DT
   • gridExtra
   • corrplot
   • glue
   • scales
   • broom

3. Option 1: Run the following scripts (in order) in terminal from the main repo directory with the specified arguments:

   a) Load data Rscript scripts/load.R --data_url=https://raw.githubusercontent.com/STAT547-UBC-2019-20/data_sets/master/listings-Barcelona.csv

   b) Clean data Rscript scripts/process.R --path_raw=data/raw_listings.csv --path_clean=data/clean_listings.csv

   c) Exploratory data analysis Rscript scripts/EDA.R --path_clean=data/clean_listings.csv --path_image=images/

   d) Linear Regression Rscript scripts/lm.R --path_data=data/clean_listings.csv

   e) Knit final report Rscript scripts/knit.R --final_report="docs/final_report.Rmd"

4. Option 2: Run make in terminal from the main repo directory to run all individual scripts above:

   a) Dependency Ensure make is installed

   b) Run all scripts and reproduce analysis make all

   c) Delete all output from scripts make clean

4. Dashboard Proposal

Dashboard Description

This app will have features for data exploration and data analysis. Data exploration uses a density of Airbnb listing prices for Barcelona, Spain and allows the user to select a categorical variable to view a violin plot. The default density is for all listings. Grouping options include District, Room Type,and Minimum Night Stay.

For data analysis, users can filter the dataset using price, distance from city center, or minimum night stay. Then, a simple linear regression is run using an independent variable (x-axis variable) chosen by the user against the dependent variable price. Data transformations: logarithms, square root, or reciprocal are available for either variable which may improve the linear relationship. Finally, a scatterplot with the linear regression line is presented.

Usage Scenarios

Maria lives in Barcelona and wants to rent out her apartment while she travels for two weeks. She decides to use Airbnb.com, and her friend has offered to manage the rentals while she is away. Maria wants a listing price that is competitive, but not too low. She knows her apartment is in a very desirable neighourhood for tourists. To guide in the price selection, she wants to compare her apartment to similar Airbnb listings. To maximize occupancy, she wants to write a description emphasizing characteristics strongly related to higher listing prices. Maria visits our dashboard and uses the data exploration section to see the distribution of listing prices for her neighborhood. To see the average listing price for comparable apartments, she uses the data analysis section to first eliminates noncomparables and outliers. Then, she selects an x-axis variable and can use her apartment's characteristics to see the mean listing price for similar listings. She decides to charge 10% over the mean. By repeating this process for a few different independent variables, she arrives at a reasonable price and knows which characteristics to stress in her listing.

Dashboard Sketch

Dashboard Usage

There are two ways to view the dashboard.

1. Click on this link to run the deployed dashboard on heroku.

2. Run the dashboard locally by taking the following steps:

   • Clone this repository locally

   • From the main repo directory in Terminal, run the dashboard by entering the command: Rscript app.R

   • To view the dashboard:

     • copy the number sequence of the form: 127.0.0.1:8050 that appear after start: in the terminal

     • paste the number sequence into the address bar of a web browser such as chrome or firefox