Final Projects

Integrating Technical Skills with Critical Thinking

1 FIN510 Final Projects

The final project is your opportunity to demonstrate both technical competence and critical thinking in financial data science. Rather than simply showcasing technical skills, these projects emphasize thoughtful application of methods, honest assessment of limitations, and integration of both traditional and causal reasoning approaches.

Choose from four project options that reflect real-world financial challenges while maintaining appropriate intellectual humility about what our analyses can and cannot tell us.

1.1 Project Options (Choose One)

1.1.1 🤖 Option 1: Causally-Informed Trading Strategy

Develop a trading system that integrates both traditional ML and causal reasoning.

Key Components: - Predictive Models: Traditional ML approaches for price prediction with honest assessment of accuracy - Causal Analysis: Investigation of whether identified relationships reflect true causal effects - Backtesting Framework: Realistic constraints with discussion of potential biases and limitations - Risk Management: Position sizing with acknowledgment of model uncertainty - Critical Evaluation: Honest discussion of when the strategy might fail and why

Required Analysis: - Compare correlation-based vs. causally-informed features - Discuss assumptions about market efficiency and regime stability - Address potential survivorship bias and look-ahead bias - Evaluate strategy robustness across different market conditions

1.1.2 📊 Option 2: Financial NLP with Causal Awareness

Develop a sentiment analysis system that thoughtfully explores text-market relationships.

Key Components: - Multi-source Analysis: News and social media sentiment with acknowledgment of source reliability issues - Causal Investigation: Explore whether sentiment influences markets or reflects underlying conditions - Temporal Analysis: Investigate timing relationships and market efficiency - Critical Assessment: Honest evaluation of when sentiment analysis is useful and when it isn’t

Required Analysis: - Compare different sentiment analysis methods and their limitations - Investigate potential confounding factors (market conditions, company fundamentals) - Discuss the direction of causality: does sentiment drive prices or vice versa? - Address challenges of real-time implementation and market efficiency

1.1.3 💼 Option 3: Robo-Advisory with Causal Risk Assessment

Develop a portfolio management system that thoughtfully considers causal relationships in risk modeling.

Key Components: - Risk Profiling: Questionnaire and algorithm with acknowledgment of behavioral complexity - Portfolio Optimization: Modern Portfolio Theory with discussion of assumptions and limitations - Causal Risk Analysis: Investigate whether correlations reflect true risk relationships - Uncertainty Quantification: Communicate confidence intervals and model limitations to hypothetical clients

Required Analysis: - Compare traditional correlation-based risk models with causally-informed approaches - Discuss assumptions about return distributions and market efficiency - Address challenges of dynamic rebalancing and transaction costs - Evaluate how recommendations might change under different market regimes

1.1.4 🎯 Option 4: Credit Risk Assessment with Fairness Analysis

Build a credit scoring system that addresses both predictive accuracy and algorithmic fairness.

Key Components: - Feature Engineering: Traditional and alternative data with discussion of causal relevance - Model Comparison: Multiple ML approaches with honest assessment of trade-offs - Bias Detection: Statistical and causal approaches to identifying unfair discrimination - Governance Framework: Model monitoring with acknowledgment of evolving fairness standards

Required Analysis: - Investigate whether predictive features reflect causal relationships with creditworthiness - Compare different fairness metrics and their philosophical foundations - Discuss challenges of balancing accuracy with fairness - Address limitations of bias detection methods

1.2 Project Requirements

1.2.1 Technical Requirements

  • Implementation: Complete Python application with clear acknowledgment of limitations
  • Documentation: Transparent code documentation including assumptions and constraints
  • Testing: Unit tests and validation with consideration of edge cases and failure modes
  • Deployment: Practical implementation with appropriate disclaimers about model uncertainty

1.2.2 Analytical Requirements

  • Traditional Analysis: Implementation of established methods with proper statistical rigor
  • Causal Investigation: Exploration of causal relationships where appropriate, with honest assessment of what can be concluded
  • Comparative Approach: Comparison of different methods with discussion of trade-offs
  • Critical Evaluation: Thoughtful assessment of limitations and potential failure modes

1.2.3 Report Requirements

  • Length: 2,000 words maximum
  • Format: Professional report balancing technical detail with accessible explanation
  • Required Sections:
    • Executive summary with clear statement of limitations
    • Methodology with transparent discussion of assumptions
    • Results with appropriate uncertainty quantification
    • Critical analysis of what conclusions are and are not supported
    • Honest discussion of future research needs
  • References: Mix of academic sources and industry applications

1.2.4 Submission Details

  • Due Date: January 14th, 2026 by 12:00
  • Format: Word document via Turnitin + GitHub repository
  • File Naming: SurnameFirstNameBNumber_ProjectTitle

1.3 Getting Started

1.3.1 1. Choose Your Project

  • Review all four options carefully
  • Consider your career interests and strengths
  • Discuss with instructor during office hours

1.3.3 3. Development Timeline

  • Weeks 9-10: Project planning and initial development
  • Week 11: Intensive development and testing
  • Week 12: Final implementation and report writing
  • Week 13-15: Final polishing and submission

1.4 Assessment Criteria

See Assessment Overview for detailed rubrics.

1.5 Project Support

1.5.1 Resources Available

  • Code Templates: Starter repositories for each project type
  • Sample Data: Curated datasets for development
  • Documentation: Technical guides and best practices
  • Examples: Anonymized high-quality past projects

1.5.2 Getting Help

  • Weekly Consultations: Individual progress meetings
  • Peer Review Sessions: Structured feedback opportunities
  • Technical Support: Lab assistants for coding issues
  • Writing Support: University academic writing services

Ready to start? Check out the project templates on GitHub.