Spot Workplace Discrimination Patterns with AI

Spot Workplace Discrimination Patterns with AI analyzes employee reviews from Glassdoor to identify disparities in workplace experiences among different demographic groups. By leveraging AI and data visualization, it calculates key metrics such as average ratings, z-scores, and effect sizes, highlighting significant gaps in employee satisfaction. This workflow empowers organizations to address inequities and foster a more inclusive workplace environment.

7/4/2025
38 nodes
Complex
76eyz9x3gu4ptgssey2mx4vnav8ckvaomanualcomplexlangchainquickchartsticky noteadvancedapiintegration
Categories:
Manual TriggeredComplex Workflow
Integrations:
LangChainQuickChartSticky Note

Target Audience

This workflow is designed for:
- HR Professionals: To analyze workplace discrimination patterns based on employee demographics.
- Data Analysts: To extract and interpret data from Glassdoor reviews to identify biases.
- Diversity and Inclusion Officers: To monitor workplace equity and assess the effectiveness of diversity initiatives.
- Company Executives: To gain insights into employee experiences and make informed decisions to improve workplace culture.

Problem Solved

This workflow addresses the issue of workplace discrimination by analyzing employee reviews from Glassdoor. It identifies disparities in ratings across various demographic groups, providing insights into potential biases and areas for improvement. By quantifying employee experiences, organizations can take actionable steps to foster a more inclusive environment.

Workflow Steps

  • Manual Trigger: The workflow is initiated by clicking 'Test workflow'.
    2. Set Company Name: It sets the target company name (e.g., Twilio) for data extraction.
    3. Define Demographic Keys: Establishes a dictionary of demographic identifiers (e.g., Asian, Hispanic, Female).
    4. ScrapingBee Search: Utilizes ScrapingBee to fetch relevant data from Glassdoor based on the company name.
    5. Extract Company URL: Extracts the URL path of the company page from the search results.
    6. Fetch Company Page: Retrieves the company page content from Glassdoor.
    7. Extract Reviews URL: Captures the URL for the reviews section of the company page.
    8. Fetch Reviews Content: Gathers reviews data from the extracted URL.
    9. Extract Overall Review Summary: Pulls the overall review summary for analysis.
    10. Extract Demographics Module: Captures demographic information from the reviews.
    11. Calculate Ratings: Analyzes ratings and their distributions using AI models to extract average ratings and review counts by demographic.
    12. Calculate Variance and Standard Deviation: Computes variance and standard deviation based on review distributions to understand rating spread.
    13. Calculate Z-Scores and Effect Sizes: Determines z-scores for demographic groups to assess how their ratings compare to the overall average.
    14. Calculate P-Scores: Evaluates the statistical significance of differences between groups.
    15. Format Data for Visualization: Prepares the dataset for visualization in scatterplots and bar charts.
    16. Generate Charts: Uses QuickChart to create visual representations of the data, highlighting disparities and trends.
  • Customization Guide

    Users can customize this workflow by:
    - Changing the Company Name: Modify the company_name variable to analyze different organizations.
    - Adjusting Demographic Keys: Edit the demographic keys in the Define dictionary of demographic keys step to fit specific needs or add new categories.
    - Modifying ScrapingBee Credentials: Update the credentials in the ScrapingBee Search and ScrapingBee GET nodes for data access.
    - Fine-tuning AI Parameters: Adjust the AI model parameters in the OpenAI nodes to refine data extraction and analysis results.
    - Altering Visualization Options: Customize the chart options in the Specify additional parameters for scatterplot step to change the appearance of the generated visualizations.