Inspiration
Most time trackers are just digital stopwatches—they tell you what you did, but not how well you did it. We were inspired by the concept of "Energy Management over Time Management." We wanted to build a tool that doesn't just shame users for losing time, but actually coaches them to align their most difficult tasks with their peak internal energy levels.
What it does
PeakPulse AI is an intelligent productivity auditor.
Smart Tracking: Features a live execution timer and a manual log with built-in validation to prevent data errors.
The Eisenhower Audit: Automatically sorts tasks into four quadrants (Do, Schedule, Delegate, Eliminate) to show users where they are "fighting fires" vs. doing deep work.
Shutterstock Energy Correlation: Uses a scatter-plot analysis to map task duration against the user's energy levels (1-5), identifying "Circadian Mismatches."
AI Coaching: Generates actionable insights, such as warning users when they are attempting productive work during low-energy "slump" periods.
The Pomodoro Engine: A built-in focus tool to help users break through "90-minute focus fatigue."
How we built it
We utilized Python for the core logic and Streamlit for the frontend to ensure a fast, reactive user experience. Data is managed using Pandas for real-time aggregation and metric calculations. We implemented a custom Session State architecture to handle the "stateless" nature of Streamlit, allowing a live timer to run without losing data on page refreshes.
Challenges we ran into
The biggest hurdle was State Management. Because Streamlit scripts rerun from the top on every interaction, keeping a live timer "alive" while the user navigated other tabs required a complex session-state logic. We also had to solve for data integrity, specifically handling edge cases like "negative time" entries and overlapping task logs.
Accomplishments that we're proud of
Logic Integration: We successfully turned a simple table of numbers into a strategic Eisenhower Matrix.
Visual Analytics: We built a dashboard that provides a "Focus Score" at a glance, making productivity feel like a game.
User Experience: We created a clean, intuitive UI that balances heavy data analysis with a simple, distraction-free tracker.
What we learned
We learned that data without context is useless. Knowing you spent 3 hours on "Email" doesn't help unless you know that those 3 hours happened during your peak energy window when you should have been coding. We also deepened our understanding of building persistent applications within the Streamlit framework.
What's next for PeakPulse AI
Database Persistence: Moving from session-based storage to a SQL/Supabase backend so users can track progress over months.
Automated Categorization: Using Natural Language Processing (NLP) to automatically sort tasks into quadrants based on the task name.
Calendar Integration: Syncing with Google Calendar or Outlook to automatically import meetings into the time audit.
Log in or sign up for Devpost to join the conversation.