Vocation Tangent

๐Ÿ”— Live site: https://hacklyticscareerlens.vercel.app/

Why we built this

Choosing a career is a big decision, but most advice online is based on opinions, trends, or vague claims. At the same time, the U.S. government publishes decades of real job market data that most people never look at because itโ€™s hard to understand. We wanted to take that real data and turn it into something visual, interactive, and actually useful.

What it does

You pick a job. The website shows how that career has changed over time, compares across states, how risky it is to automation, and what current trends suggest for the future. Everything you see is based on real historical data, not scraped job postings or generated numbers.

The data

All the numbers in Vocation Tangent come from official U.S. labor datasets, mainly:

  • Bureau of Labor Statistics (Occupational Employment Statistics)
  • O*NET job and work context data
  • Salaries are reported wages, not estimates from job ads
  • Demand is based on actual employment numbers
  • Trends are consistent and comparable across many years

Gemini AI is only used to analyze and explain this data,it does not replace it.

Main features

24+Year Heatmap

A simple way to see how a career changes year by year across salary, demand, stability, and automation risk.

Tradeoff Comparison

A visual comparison of up to three careers to help see real tradeoffs instead of vague pros and cons.

Ask Gemini AI

You can ask questions like โ€œWhy did nurse salaries spike in 2020?โ€ and get a clear explanation powered by Gemini API, based on the actual data.

How we built it

Frontend & Visualization Vanilla HTML, CSS, JavaScript (no framework) D3.js for the U.S. map Chart.js for radar charts and forecasts Linear Regression to capture long-term trends Random Forest to handle non-linear patterns and fluctuations

Designed to be fully interactive, every hover and dropdown is live

What makes this different

  • Uses government labor data, not scraped listings
  • Clearly separates historical facts from projections
  • Uses Actian VectorAI as a core part of the analysis
  • Fully interactive and deployed, not a static demo

Built With

  • claude
  • fastapi
  • geminiapi
  • python
  • vectordb
Share this project:

Updates