Spot Workplace Discrimination Patterns with AI

Workflow Name

Spot Workplace Discrimination Patterns with AI

Key Features and Highlights

This workflow automates the extraction and analysis of employee review data from Glassdoor using ScrapingBee to bypass JavaScript restrictions and efficiently obtain raw data. It then leverages OpenAI’s powerful language models to perform in-depth analysis of overall company ratings and rating distributions across different demographic groups. Statistical metrics such as Z-scores, effect sizes, and p-values are calculated to quantify differences. Visualizations including scatter plots and bar charts are generated via QuickChart to clearly illustrate disparities in workplace experiences among various groups.

Core Problems Addressed

Traditional workplace discrimination patterns are difficult to quantify and visualize, and anonymous employee feedback often fails to directly reflect true group differences. This workflow automates the collection and analysis of large volumes of authentic employee reviews, enabling HR and management to detect and quantify potential workplace discrimination and inequalities. It supports data-driven initiatives for fair and equitable improvements.

Application Scenarios

• Employee satisfaction and diversity equity analysis by corporate HR departments
• Organizational culture enhancement and inclusivity assessment
• Formulation and monitoring of anti-discrimination policies
• Data-driven support for Diversity, Equity, and Inclusion (DEI) programs
• Academic research and industry report data collection and analysis

Main Process Steps

1. Manually trigger the workflow to start analysis.
2. Set the target company name (default is Twilio, customizable).
3. Use the ScrapingBee API to access Glassdoor’s search page and obtain the target company’s page path.
4. Request and scrape the company’s Glassdoor homepage and review pages.
5. Extract overall review summaries and demographic module HTML content.
6. Utilize OpenAI models to parse overall rating distributions, average ratings, and review counts by demographic groups.
7. Calculate variance and standard deviation of rating distributions.
8. Compute Z-scores, effect sizes, and p-values for each demographic group relative to the overall population to assess significance of differences.
9. Format the calculated data sets and generate visual scatter plots and bar charts.
10. Use OpenAI to generate textual summaries of the analysis, highlighting key insights and descriptions of employee experiences.

Involved Systems and Services

• ScrapingBee (web data scraping proxy service)
• Glassdoor (source of anonymous employee review data)
• OpenAI (natural language processing and information extraction)
• QuickChart (chart generation and visualization)
• n8n (workflow automation platform)

Target Users and Value Proposition

• Corporate HR and Diversity, Equity & Inclusion (DEI) teams seeking to identify and address workplace discrimination
• Organizational leaders aiming to improve employee experience and culture based on data insights
• Data analysts and researchers focused on workplace group disparities and equity studies
• Consulting firms providing workplace fairness assessment services to clients
• Professionals leveraging automated tools to gain deep understanding of workplace review data for equitable management

---

By combining automation, data-driven methodology, and advanced AI analytical capabilities, this workflow empowers organizations to uncover hidden biases and discrimination patterns in employee reviews, fostering a fairer and more inclusive workplace environment.