Gubli A.S

Case study

Brimwatch

A smart drain monitoring and flood prevention system for Kumasi. A sensor in the drain warns the city before it floods, and residents can report a blocked drain from their phone in seconds.

Context
KNUST Computer Science project
Team
Three people
My role
Built the website and platform, helped build the hardware
Year
2026

Tools used

  • ESP32
  • Next.js
  • React
  • Tailwind CSS
  • Three.js
  • Express
  • Prisma
  • PostgreSQL
  • MapLibre
The working prototype: sensors, board and live test
The working prototype: sensors, board and live test
Report a blocked drain in three steps
Report a blocked drain in three steps
The city's live map and work list
The city's live map and work list

Project overview

Flooding from blocked drains happens every rainy season in Ghanaian cities, and it is largely preventable. In Kumasi, drains are checked on a fixed schedule or only after a flood, so nobody knows a drain is blocked until water is already in the streets.

Brimwatch gives the city that missing view. A sensor inside the drain measures how full it is, residents can report a blocked drain with a photo, and the city sees both on one live map. When a drain reaches a dangerous level, the right people get a text message straight away.

  • A sensor that measures how full a drain is and senses when it is raining
  • A reporting page any resident can use from their phone: a photo, the location found automatically, done
  • A live map for the city's maintenance team, coloured by how serious each drain is
  • Text message warnings to the city and disaster teams when a drain becomes critical
  • A list of the most urgent drains, so crews go where they are needed first

The challenge

Sensors are reliable but there can only ever be so many of them. Residents are everywhere but only report what they notice. Relying on either one alone leaves gaps,.

  • Spot blockages before the rain turns them into floods
  • Make reporting so easy that people actually do it

Approach

We combined both: sensors for constant readings, and residents for reach across the whole city. The two show up on the same map, so the system is useful from day one.

I built the whole website and platform: the public site, the reporting page for residents, and the live map for the city team. I also helped my teammates build the sensor. We proved it worked live: as the water rose, the map changed and the warning went out.

  • Sensor readings and residents' reports on one map
  • Reporting in a few taps, straight from a phone
  • Warnings by text, so nobody has to be watching a screen
A drain fills, unseen
A drain fills, unseen
The city responds
The city responds
The cost of flooding
The cost of flooding
The website on a phone
The website on a phone
The parts, before assembly
The parts, before assembly
Testing the sensor with live readings
Testing the sensor with live readings

What this shows

A working prototype of the full journey, from a drain filling up to the right people being warned, shown live with a real sensor.

I can build the software for a hardware product and help with the hardware itself, turning raw sensor readings into a clear map, warnings and tools that city staff and ordinary residents can use.

  • Built the full website and platform myself
  • Designed for real users: residents and city crews
  • Helped build and test the hardware