Knowledge Sharing

Documenting the 3MTT Capstone journey, technical implementations, and community impact.

The 3MTT Knowledge Showcase

Aligning with the core webinar mandates for Cohort 2 project presentation.

The WasteWise AI project is explicitly designed not just as a software application, but as a transparent, open-learning ecosystem to fulfill the Nigerian Federal Ministry of Communications' 3MTT Knowledge Sharing initiative.

Building this predictive infrastructure required merging insights from the AI/ML NextGen track. It involved integrating raw data structuring and applying K-Nearest Neighbors (KNN) machine learning logic to predict bin overflow based on geospatial clustering and waste decay severity.

By open-sourcing the thought process and the analytical routing logic, this project serves as a replicable template for other civic tech developers looking to optimize urban resource management.

Technical Learnings

Predictive Fallbacks: Learned how to architect serverless database constraints on Vercel by implementing an in-memory AI calculator fallback to ensure uninterrupted demonstrations.
UI/UX Rigidity: Solved complex Tailwind grid offset issues that pushed critical content off-center on ultra-wide screens by utilizing strict `mx-auto` boundary constraints.
Data Pipelines: Translated raw Python ML scripts into functional Next.js/React hooks to run local predictions dynamically.

Community Impact

Scalability: The underlying data architecture can be instantly migrated to any urban agency globally, mapping entirely new districts without altering the algorithm.
Open Source Reusability: All dashboard analytics code, React-Leaflet map implementations, and AI risk heuristics will be documented in a public repo to assist upcoming 3MTT cohorts.

Explore the Open Source Codebase

3MTT Official Evaluation Matrix

Explicit platform alignment with the 5 graded pillars established in the Capstone Webinar.

1. Area of Innovation (20%)

Transitioning urban waste management from highly-reactive complaint-based routing into predictive, AI-driven logistic triaging to intercept catastrophic public health overflows before they escalate.

2. Technical Quality (25%)

Deployed on modern edge-networks via Next.js and Vercel, integrating Python-style K-Nearest Neighbor (KNN) logic directly into TypeScript server-hooks. Features advanced data schemas and responsive layout matrices.

3. Functionality (20%)

WasteWise AI is not a wireframe. It is a live, functional multi-route MVP. It intercepts read-only deployment constraints by injecting Serverless AI memory fallbacks, ensuring the civic submission pipeline never crashes during a live demo.

4. Learning Application (20%)

Demonstrates absolute mastery over the AI/ML curriculum. Simulated mock datasets were curated, and Machine Learning heuristic models were deployed directly into the application matrix to compute spatial risks.

5. Presentation (15%)

Designed with extreme enterprise aesthetics to guarantee a persuasive 2-4 minute video demonstration, exhibiting confidence in architectural design and problem-solving transparency.