Skip to content

nerdjumbo/Glassdoor-Sentiment-Analysis

 
 

Repository files navigation

Disclaimer

This scraper is provided as a public service because Glasdoor doesn't have an API for reviews. Glassdoor TOS prohibit scraping and I make no representation that your account won't be banned if you use this program. Furthermore, should I be contacted by Glassdoor with a request to remove this repo, I will do so immediately.

Introduction

Have you ever wanted to scrape reviews from Glassdoor, but bemoaned the site's lack of a public API for reviews? Worry no more! This script will go through pages and pages of reviews and scrape review data into a tidy CSV file. Pass it a company page and set a limit to scrape the 25 most conveniently available reviews, or control options like the number of reviews to scrape and the max/min review publication date.

It takes about 1.5 seconds per review to scrape. So it will take about 25 minutes to scrape 1,000 reviews, or a little over 4 hours to scrape 10,000 reviews. This script requires patience. 😁

Installation

First, make sure that you're using Python 3.

  1. Clone or download this repository.
  2. Run pip install -r requirements.txt inside this repo. Consider doing this inside of a Python virtual environment.
  3. Install Chromedriver in the working directory.
  4. Create a secret.json file containing the keys username and password with your Glassdoor login information, or pass those arguments at the command line. Note that the second method is less secure, but in any case you should consider creating a dummy Glassdoor account.

Usage

usage: main.py [-h] [-u URL] [-f FILE] [--headless] [--username USERNAME]
               [-p PASSWORD] [-c CREDENTIALS] [-l LIMIT] [--start_from_url] 
               [--max_date MAX_DATE] [--min_date MIN_DATE]

optional arguments:
  -h, --help                                  show this help message and exit
  -u URL, --url URL                           URL of the company's Glassdoor landing page.
  -f FILE, --file FILE                        Output file.
  --headless                                  Run Chrome in headless mode.
  --username USERNAME                         Email address used to sign in to GD.
  -p PASSWORD, --password PASSWORD            Password to sign in to GD.
  -c CREDENTIALS, --credentials CREDENTIALS   Credentials file
  -l LIMIT, --limit LIMIT                     Max reviews to scrape
  --start_from_url                            Start scraping from the passed URL.
  
  --max_date MAX_DATE                         Latest review date to scrape. Only use this option
                                              with --start_from_url. You also must have sorted
                                              Glassdoor reviews ASCENDING by date.
                                              
  --min_date MIN_DATE                         Earliest review date to scrape. Only use this option
                                              with --start_from_url. You also must have sorted
                                              Glassdoor reviews DESCENDING by date.

Run the script as follows, taking Wells Fargo as an example. You can pass --headless to prevent the Chrome window from being visible, and the --limit option will limit how many reviews get scraped. The-f option specifies the output file, which defaults to glassdoor_reviews.csv.

Example 1

Suppose you want to get the top 1,000 most popular reviews for Wells Fargo. Run the command as follows:

python main.py --headless --url "https://www.glassdoor.com/Overview/Working-at-Wells-Fargo-EI_IE8876.11,22.htm" --limit 1000 -f wells_fargo_reviews.csv

Note: To be safe, always surround the URL with quotes. This only matters in the presence of a query string.

Example 2: Date Filtering

If you want to scrape all reviews in a date range, sort reviews on Glassdoor ascending/descending by date, find the page with the appropriate starting date, set the max/min date to the other end of your desired time range, and set limit to 99999.

Suppose you want to scrape all reviews from McDonald's that were posted in 2010:

  1. Navigate to McDonald's Glassdoor page and sort reviews ascending by date.
  2. Find the first page with a review from 2010, which happens to be page 13.
  3. Send the command to the script: python main.py --headless --start_from_url --limit 9999 --max_date 2010-12-31 --url "https://www.glassdoor.com/Reviews/McDonald-s-Reviews-E432_P13.htm?sort.sortType=RD&sort.ascending=true"

If there's demand for it, we can automate this process to provide a simple interface for filtering by date.

About

No description, website, or topics provided.

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages

  • Python 100.0%