This project was created with the initial purpose of learning to scrape 'messy' data and clean it through a pipeline of functions automatically. Some features may not be implemented perfectly, or be missing entirely. Clean, functional, decoupled code is the main purpose of this project, along with learning how to implement traditional relational databases and NoSQL databases.

• Scrapes all of the listings for search term
• Cleans data for analysis
• Stores data in database
• Event logging
• NLP of descriptions and amenities
• Machine learning and visualization of price influencers.

• Load different formats into database
• MongoDB integration for articles
• Visualize Data
• Machine Learning algorithms to find key price predictors.
• Options Run on Command Line
• Web Interface with more options

trulia_scrape.py - this file can be run from the command line and will automatically scrape apartment data for the Austin area and save it to a CSV file in the daily_scrape_files folder.

# Clone this repository
$ git clone https://github.com/datapointchris/etl_housing

# Go into the repository
$ cd etl_housing

# Run the app
$ python scraper.py

Program will begin scraping Trulia for rentals. Currently only Austin rentals have been tested. Other cities and search terms will be available in future versions.

Jupyter Notebooks are also included in the repo where you can run the program and change the page_url to scrape different cities.

You really should only need to install BeautifulSoup if you don't have it. Everything else should be part of the standard library.

• Numpy
• Pandas
• Requests
• BeautifulSoup
• SQLite3

MIT

# Run Redis Docker
docker run -d --name redis-server -p 6379:6379 redis
# Set up network
docker network create page-tracker-network
# Run Redis Docker on network
docker network connect page-tracker-network redis-server
# Run Redis CLI
docker run -it --network page-tracker-network --rm redis redis-cli -h redis-server