Inspiration
Economic policy decisions are complex, nonlinear, and full of tradeoffs that are difficult for everyday people to understand
Most existing tools (government dashboards, academic models) are static, slow to interpret, and not interactive
Students and citizens often can’t visualize how policy changes affect real-world outcomes like unemployment, inflation, or debt
We wanted to make economic data and forecasting more accessible, interactive, and understandable
Our goal was to build a tool that helps users explore “what happens if…” scenarios in real time
What it does
Displays live U.S. economic data (unemployment, inflation, interest rates, etc.) using public datasets
Uses time-series forecasting models to predict future trends in key economic indicators
Provides an interactive dashboard where users can toggle filters and explore different economic metrics
Helps users understand tradeoffs in policy and financial decisions through visual projections
Presents data in an accessible, easy-to-understand format rather than dense academic graphs
Designed for students, citizens, and anyone curious about economic trends
How we built it
Pulled real-time economic data from the FRED (Federal Reserve Economic Data) API
Cleaned and processed time-series data using Python, pandas, and NumPy
Built predictive models using ARIMA/Auto-ARIMA for economic forecasting
Created a backend using Flask and Node.js to serve predictions and handle API requests
Designed a web dashboard with modern frontend tools to visualize trends and forecasts
Integrated model predictions into interactive charts for real-time exploration
Challenges we ran into
Cleaning and standardizing time-series data across different frequencies (daily vs monthly)
Selecting and tuning appropriate forecasting models for economic indicators
Handling missing data and ensuring models produced realistic predictions
Connecting multiple backend technologies (Flask + Node + frontend) smoothly
Managing environment variables and API keys securely during development
Time constraints of building and integrating models within a hackathon timeframe
Accomplishments that we're proud of
Successfully built working predictive models for real-world economic indicators
Integrated live data with forecast visualizations in a functional web app
Created a clean, extensible backend architecture for adding new indicators
Transformed complex economic data into an accessible interactive experience
Demonstrated how machine learning and data science can improve public understanding of policy impacts
Delivered a working prototype within a limited hackathon timeframe
What we learned
Time-series forecasting requires careful preprocessing and validation
Economic data often behaves differently from traditional ML datasets
Good API and backend design is just as important as model accuracy
Clear data visualization dramatically improves usability and understanding
Collaboration and clear task division are critical in short development cycles
Rapid prototyping and iteration are key during hackathons
What's next for JALT — Joint Algorithmic Labor & Tax Simulator
Add more economic indicators (GDP, consumer sentiment, housing data, etc.)
Introduce scenario-based simulations for policy changes (tax rates, spending shifts)
Improve model accuracy using more advanced forecasting techniques
Build a more polished and scalable frontend dashboard
Allow users to compare multiple economic scenarios side-by-side
Deploy the app publicly for students and researchers to explore
Expand into a broader civic-tech platform for policy education and transparency
Log in or sign up for Devpost to join the conversation.