The Silent Burnout: Mental Health in Tech

• 

  thumbnail

Inspiration

The tech industry prides itself on innovation, but often at the cost of its workers' well-being. We wanted to look past the stereotype of the "burnt out developer" and use data to find the root causes. Is it crunch time? Lack of benefits? Or simply the stigma of speaking up?

What it does

"The Silent Burnout" is an interactive data story built in Hex that analyzes over 1,200 survey responses. It features:

1. AI-Powered Data Cleaning: Standardizes messy free-text gender data into inclusive categories.
2. Interactive Exploration: Allows users to filter mental health trends by Country to see regional differences.
3. Predictive Modeling: Uses a Random Forest Classifier (Accuracy: 64.1%) to identify the strongest predictors of whether an employee will seek treatment.

How we built it

We started with raw Python scripts (pandas, scikit-learn) to define our cleaning pipeline and EDA. We then migrated this logic into Hex, leveraging its "Magic" AI to generate initial SQL/Python visualizations. The interactive dashboard was built using Hex's App builder, connecting dropdown parameters directly to our charts.

Challenges we ran into

The "Gender" column was a mess of over 50 different free-text responses (e.g., "Male-ish", "Cis Male", "woman"). Building a robust cleaning function to categorize these identities properly was our first major hurdle.

Accomplishments that we're proud of

We moved beyond simple charts to a predictive model that offered a counter-intuitive insight: The biggest driver for seeking help isn't the severity of work interference, but the availability of care options and benefits. Policy matters more than individual resilience.

Built With

• hex
• pandas
• python
• scikit-learn