Analysis on data scraped from Glassdoor

We scrape data from Glassdor first and then do some interesting analysis.

Requirements

• Chrome v78
• Python3
• Python packages: bs4, selenium, unicodecsv, pandas, numpy, nltk
• Chromedriver from selenium webpage. https://chromedriver.chromium.org/downloads

Scraping

1. Steps to scrape salary data:

• First change director by cd Scraper/SalaryScraper and edit file salary_scraper_specific.py to include your Glassdor username and password.
• Now run python salary_scraper_specific.py and wait till scraper stops.

2. Steps to scrape review data:

• First change director by cd Scraper/ReviewScraper, and then run command python run_all.py.

Data pre-processing

After we obtain review and salary data for each company, the next step is to merge these individual tables and pre-process to clean the outliers. This stage generate following tables:

• merge_reviews_table.csv
• fulltime_merged_salaries_company_table.csv
• intern_merged_salaries_company_table.csv

In order to generate above tables run python merge_table.py

Analysis

Our analysis procedure tries to answer following questions.

• Which company offers the highest average salary?
• Which field has more jobs?
• Are interns paid generously?
• Which job category provide the highest average salary?
• Which state in US offers the most job opportunities?
• Which city should you move if you are looking for a job?
• How do employees rate their CEOs?
• Will they recommend their company to their friends?
• What feedback do employees give for the companies they are working?

These analysis can be seen by running demo.ipynb notebook.