Due to javascript restrictions, a normal HTTP request cannot be used to gather user-reported details from Glassdoor.

Instead, ScrapingBee offers a great tool with a very generous package of free tokens per month, which works out to roughly 4-5 runs of this workflow.

How this workflow works

1. Replace ScrapingBee and OpenAI credentials
2. Replace company_name with company of choice (workflow performs better with larger US-based organizations)
3. Preview QuickChart data visualizations and AI data analysis

Inspired by Wes Medford's Medium Post

Wes performed the initial data analysis highlighting problematic behaviors at Twilio. I wanted to try and democratize the data analysis they performed for those less technical.

Hi, Wes!

Glossary

Z-Score – A statistical measure that indicates how many standard deviations a data point is from the mean. In this analysis, a negative z-score suggests a group rates their workplace experience lower than the average, while a positive z-score suggests a better-than-average experience.

Effect Size – A measure of the magnitude of difference between groups. Larger negative effect sizes indicate a more substantial disparity in workplace experiences for certain groups, making it useful for identifying meaningful gaps beyond just statistical significance.

P-Score (P-Value) – The probability that the observed differences occurred by chance. A lower p-score (typically below 0.05) suggests the difference is statistically significant and unlikely to be random. In this analysis, high p-scores confirm that the disparities in ratings for marginalized groups are unlikely to be due to chance alone.

Relevance to This Analysis

These metrics help quantify workplace disparities among demographic groups. Z-scores show which groups report better or worse experiences, effect sizes reveal the severity of these differences, and p-scores confirm whether the disparities are statistically meaningful. This data allows for a more informed discussion about workplace equity and areas needing improvement.

Example AI Analysis (Twilio Example)

Key Takeaways

1. Significant Disparity Among Disabled Employees Disabled employees reported the lowest average ratings, with a z-score of -5.50, indicating a far worse experience compared to their non-disabled peers.
2. LGBTQIA Community's Challenges Members of the LGBTQIA community showed significantly lower ratings (z-score of -2.53), suggesting they may experience a workplace environment that is less inclusive or supportive compared to others.
3. Transgender Experiences Are Particularly Negative Transgender employees rated their experiences considerably lower (z-score of -2.91), highlighting a critical area for improvement in workplace culture and acceptance.
4. Veterans Report Higher Satisfaction In contrast, veterans had the highest ratings (z-score of 1.54), which could indicate a supportive environment or programs tailored to their needs.
5. Overall Gender Discrepancies A noticeable gap exists in average ratings by gender, with female employees scoring below male employees, suggesting potential gender biases or challenges in workplace dynamics.

Employee Experiences

Perceptions of Workplace Environment

For members of groups reporting significantly worse experiences, such as disabled, transgender, and LGBTQIA employees, the workplace may feel alienating or unwelcoming. These individuals might perceive that their contributions are undervalued or overlooked and that necessary support systems are lacking, creating a culture of exclusion rather than one of inclusivity. This feeling of being marginalized can lead to poorer engagement, higher turnover rates, and diminished overall job satisfaction, adversely impacting both employees and the organization.

Spot Workplace Discrimination Patterns with AI

Purpose and structure

This folder contains the Spot Workplace Discrimination Patterns with AI workflow. Its executable nodes use chainLlm, code, html, httpRequest, informationExtractor, lmChatOpenAi, manualTrigger, merge, quickChart, set. The table and connection map below are derived from the JSON, not an inferred guarantee of behavior.

Setup

Import inactive. Configure your own credentials and replace REPLACE_WITH_* values. Review external destinations and Code nodes before running.

Entry points

When clicking ‘Test workflow’

Verification

Sanitized source candidate; not runtime-tested. This overview is generated from node metadata. Read the folder README for the connection map, credential types, and license.

Attribution

https://github.com/Zie619/n8n-workflows/blob/94007c1445d9258a7da116646b79473e7c7c3282/workflows/Code/1994_Code_Manual_Automation_Webhook.json MIT; preserve source notices.