FarmLink Intelligence is a final-year OOAD project by eight Computer Science students at The ICT University — combining software engineering and IoT hardware to solve real problems in Cameroonian agriculture.
Agriculture accounts for over 20% of Cameroon's GDP. Yet smallholder farmers — the backbone of the sector — sell to middlemen at a fraction of the real market price, because buyers have no way to trust quality claims made online.
FarmLink Intelligence changes this by attaching real IoT sensor data directly to every marketplace listing. Quality is no longer a claim — it is a measurement. Trust is no longer assumed — it is computed from every transaction.
We built this platform as engineering students who see this problem every day. Our goal was never to build another marketplace. Our goal was to build a supply intelligence engine.
Every quality score shows its sensor source. Buyers never see a naked number without understanding where it came from.
Farmer reliability is earned through real transaction history — not self-reported. Honest farmers gain a visible competitive advantage.
Every complex action is broken into the smallest possible steps. If a farmer with limited tech experience cannot complete it in 3 minutes, the design has failed.
Anything the system computes automatically — quality score, trust score, harvest forecast — is read-only. Farmers cannot edit what sensors measure.
"Don't build a marketplace. Build a supply intelligence engine that connects sensor data, trust reputation, and predictive forecasting into a single farm-to-buyer pipeline."
— FarmLink Intelligence design principle, v1.0These three problems exist simultaneously in Cameroonian agriculture. FarmLink Intelligence is the first platform designed to solve all three at once.
Cameroonian smallholder farmers have no direct channel to urban buyers. They sell through intermediaries at a fraction of the real price — not because their produce is inferior, but because they are invisible in the market. Buyers, meanwhile, have no way to reach farmers directly, so they rely on middlemen who inflate costs.
FarmLink solution: Direct farmer-to-buyer marketplace with verified listingsAny farmer can type "high quality produce" in a listing. Without hardware evidence, buyers have no way to know if this is true. The result is a trust deficit that prevents online agricultural commerce from functioning. Even genuinely high-quality farmers cannot differentiate themselves from dishonest ones.
FarmLink solution: IoT sensor nodes generate read-only, tamper-proof quality scoresBuyers cannot plan procurement ahead of time because no platform tells them when a harvest will be ready. This leads to last-minute, reactive purchasing — driving up prices, increasing waste, and making supply chains fragile. Farmers also lose out because buyers are not waiting for them.
FarmLink solution: 28-day harvest forecast engine with automatic buyer alertsEight Computer Science students at The ICT University — two teams working in parallel on software and hardware.
Leads the FarmLink Intelligence project end-to-end — architecting the Flask backend, designing the database schema, and coordinating the full team. Responsible for the application factory, blueprint structure, and Flask-Login authentication system.
Builds the core server-side logic powering FarmLink — from API route handling to sensor data ingestion. Works closely with the IoT team to ensure Firebase readings arrive cleanly into MySQL and trigger the quality scoring pipeline.
Designs and maintains the relational database at the heart of FarmLink — defining every table, relationship, and constraint that connects users, farms, sensor readings, listings, and alerts into a single coherent data model.
Translates wireframes into pixel-perfect, responsive HTML and CSS across the FarmLink portal pages. Ensures every interface follows the design specification — from colour variables and typography to component reuse and mobile responsiveness.
Crafts the user-facing pages of FarmLink Intelligence — building responsive, accessible interfaces that work seamlessly across devices. Focuses on delivering a clean, intuitive experience for both farmers and buyers navigating the platform.
Assembles and calibrates the physical sensor nodes deployed on Cameroonian farms — integrating the DHT22, soil moisture sensor, BH1750 light sensor, and DS3231 RTC onto the Arduino Uno board and validating readings for accuracy in the field.
Builds and maintains the real-time data bridge between the ESP8266 Wi-Fi module and Firebase Realtime Database — ensuring every sensor reading is transmitted, stored, and then synced reliably into the MySQL application database.
Writes the C++ firmware running on each Arduino node — programming the 30-minute reading loop, managing the DS3231 real-time clock for accurate timestamps, and handling offline buffering so no readings are lost during connectivity gaps.
Registration is free for both farmers and buyers. No middlemen. No hidden fees.