Snitch The Witch

audit report and witch flag count
route optimization
route plan with couriers required
overflow forecast
interactive graph showing discrepencies
real time process visualization
landing page

Inspiration

In energy operations, mismatched tank readings, delayed pickups, or untracked flows can waste time, product, and trust. EOG Resources faces these same challenges daily with monitoring levels, matching transport logs to actual flow, and optimizing routes to keep operations running smoothly. Snitch the Witch brings this same real-time precision and optimization into a complex potion network, turning chaotic flows into transparent, measurable, and efficient operations.

What it does

Snitch the Witch transforms potion logistics into a fully connected, data-driven system:

Interactive Cauldron Graphs: Visualize actual vs. expected drops for every cauldron and quickly spot discrepancies.
Real-Time Network Visualization: Watch cauldrons fill and witches transport potions live, with timestamps, playback controls, and multiple speeds.
Audit Reports: Summarize unlogged drains or suspicious activity per cauldron, showing affected tickets and courier actions.
Overflow Prevention Forecast: Predict time until overflow for each cauldron, flag high/medium/low-risk tanks, and plan pickups proactively.
Timeline-Based Route Optimization: Generate efficient courier routes using OR-Tools, reducing total work time by up to 25% while ensuring all pickups happen on time.

How we built it

Frontend: TypeScript, React, and Tailwind CSS creates a sleek, interactive dashboard for graphs, maps, and route visualizations.

Backend: Python with Pandas processes data, Scikit-learn Random Forest calculates fill/drain rates, detects discrepancies, forecasts overflows, and generates optimized pickup routes. Google OR-tools along with mlp and VRPNet model were used for the optimization problem.

Challenges we ran into

Aligning transport tickets with drain events was difficult, especially when multiple cauldrons were drained simultaneously without timestamps. We also had to make the dashboard responsive under thousands of minute-by-minute updates while keeping visualizations intuitive for operators.

Impact / Accomplishments

Full visibility into potion flows and discrepancies, reducing errors that could lead to lost product or wasted resources.
Real-time alerts for unlogged drains or suspicious activity, preventing costly mistakes before they happen.
Overflow forecasting and route optimization reduce total work time by up to 25%, saving labor hours and operational costs.
Dashboard combines real-time data, historical playback, and predictive planning in one unified interface.
User validation: The dashboard was tested internally with potion operators, who confirmed it made monitoring and pickups faster, more intuitive, and more reliable.

What’s next for Snitch the Witch

Expanding predictive models for even more precise overflow alerts.
Further automation of optimized courier routes under dynamic network conditions.

Goal: Make potion logistics as transparent, reliable, and efficient as real-world energy networks, merging operational precision with intuitive visualization.