Inspiration
Traditional financial apps treat money management as a cold, purely mathematical equation. When users experience high stress, burnout, or anxiety, they frequently turn to retail therapy as an emotional coping mechanism—only to be met with punitive red alerts and guilt-inducing notifications by standard budgeting tools.
We were inspired to bridge this gap by recognizing a fundamental truth: financial health is mental health. We wanted to build a platform that meets users with empathy during moments of high cognitive load, shifting the paradigm from financial surveillance to compassionate financial agency.
What it does
Nudge is an empathetic behavioral fintech platform that fuses subjective mental health telemetry with real-time banking data. During daily check-ins, users log subjective 1–10 scores across metrics like stress, anxiety, and impulse. The engine calculates a real-time Cognitive Load Index ($CLI$) and cross-references it with recent discretionary spending.
Depending on the user's behavioral archetype and emotional state, Nudge automatically determines the Next Best Action (NBA)—ranging from executing a cooling-off savings sweep and micro-debt payoffs via a mocked open banking API to delivering comforting, non-punitive behavioral guidance.
How we built it
We designed a modular, zero-config Python backend optimized for transparent execution and rapid prototyping:
- Persistence Layer: Dual SQLite databases (
nudge_telemetry.dbandnudge_actions.db) to safely isolate and archive user check-ins and automated financial action logs. - Data Fusion & Engine: Python and Pandas process synthetic transaction logs and compute the composite Cognitive Load Index.
- API Simulation: A mocked open banking execution layer simulates secure transfer endpoints (such as Plaid or Finicity).
- Dashboard Aggregator: Compiles mental health streaks, total savings swept, and debt payoff metrics into a cohesive executive health dashboard.
Challenges we ran into
- Temporal Alignment: Ensuring that emotional telemetry matched the appropriate transaction window was critical; we had to prevent stale psychological check-ins from triggering automated financial sweeps hours later.
- Calibrating the Decision Matrix: Tuning the rules for the Cognitive Load Index ($CLI \ge 70.0$) required careful iteration to ensure the system accurately distinguished between normal daily friction and genuine impulse-spending vulnerability without being overly intrusive.
Accomplishments that we're proud of
- Successfully conceptualizing and coding a fully functional multi-modal pipeline that links qualitative psychological metrics directly to quantitative banking outcomes.
- Developing a clean, reproducible architecture contained within a single master notebook workflow, making it effortless for judges and developers to test end-to-end.
What we learned
We learned that building effective financial technology requires deeply understanding human psychology rather than just focusing on ledger arithmetic. Designing systems that acknowledge emotional exhaustion and provide gentle friction rather than strict blockades creates a far more sustainable path to long-term financial wellness.
What's next for Project Nudge
- Live Open Banking Integration: Transitioning the mocked API framework into live integrations with secure financial data networks like Plaid.
- Mobile-First Interface: Expanding the design-led wireframes into a fully realized mobile application with push-notification nudges.
- Advanced Machine Learning: Evolving the rule-based decision engine into predictive models that anticipate stress-spending patterns before they occur.
