Predictive fibre cut risk intelligence for Airtel Nigeria-3MTT Knowledge showcase 2026
# π‘ FibreWatch AI
> **Predictive Fibre Cut Risk Intelligence for Airtel Nigeria**
> 3MTT Knowledge Showcase 2026 Β· NextGen Cohort Β· FEED Pillar: **Digital Inclusion**
---
## π΄ The Problem
Airtel Nigeria faces an average of **43 fibre cuts per day**, with ~90% of damage concentrated in Abuja alone β caused by road construction, vandalism, and poor coordination between contractors and telecoms operators.
The current response is **reactive**: patrol teams are dispatched *after* the damage occurs, leaving millions of Nigerians β especially in rural and underserved communities β without connectivity for hours or days.
> **There is no proactive, data-driven system to predict where the next fibre cut will happen before it does.**
FibreWatch AI changes that.
---
## π‘ The Solution
FibreWatch AI is a **machine learning-powered risk dashboard** that:
- πΊοΈ **Clusters** historical fibre cut locations into spatial hotspots using **DBSCAN**
- π€ **Predicts** cut probability per corridor using a **Random Forest classifier**
- π **Visualises** risk zones on an interactive **Nigeria heatmap**
- π¨ **Alerts** network operations teams to HIGH-risk zones before cuts occur
- π‘ **Runs on low bandwidth** β built with Streamlit, deployable on any device
---
## π₯οΈ Dashboard Preview
```
βββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ
β π‘ FibreWatch AI | Airtel Nigeria Network Ops β
ββββββββββββββββ¬βββββββββββββββ¬βββββββββββββββ¬βββββββββββββ€
β π΄ High Risk β β οΈ Overdue β π Hotspot β π― Model β
β Zones: 47 β Patrols: 23 β Clusters: 8 β Acc: 84% β
ββββββββββββββββ΄βββββββββββββββ΄βββββββββββββββ΄βββββββββββββ€
β β
β [ INTERACTIVE NIGERIA RISK HEATMAP ] β
β π΄ High π‘ Medium π’ Low β
β β
ββββββββββββββββββββββββββββ¬βββββββββββββββββββββββββββββββ€
β π¨ Top Priority Zones β π Feature Importanc β¦