Smart SoilTrack

The Smart Soil Tracker is an offline precision farming tool that collects soil data, analyzes it, recommends crops, maps results with Mapbox, and syncs to the cloud when internet is available.

Comment

Smart Soil Tracker — Offline IoT Agriculture Platform

1. Inspiration

In many rural areas, especially in developing countries, farmers face challenges related to poor soil management, lack of access to real-time agricultural data, and the inability to make informed decisions due to the absence of reliable tools.

1.1. Problem Context

In the context of resource-constrained environments, particularly in rural agricultural communities, farmers often operate without access to modern technologies. These limitations include:

  • Limited or no internet connectivity
  • Unreliable or no electricity supply
  • Lack of real-time data for decision-making
  • Low digital literacy and technical support

These constraints hinder the implementation of conventional precision agriculture solutions, which typically rely on cloud computing, stable internet, and high-end hardware.

1.2. Inspiration for Smart SoilTrack

We were inspired to build SoilTrack as a smart, local, and sustainable solution tailored for farmers in remote and underserved regions.

Instead of relying on cloud-based systems or high-power devices, SoilTrack leverages:

  • Offline-first architecture — ensuring full functionality even without internet.
  • Low-power ESP32 microcontroller or Arduino UNO R4 WiFi — ideal for energy-constrained environments.
  • Edge-based AI processing — making smart predictions with small-data models directly on the device.
  • Local React dashboard — simple, intuitive interface for any agent or technician in the field.

2. What it does

Smart Soil Tracker is an offline-first precision agriculture platform that allows field agents to:

  • Collect real-time soil data: temperature, moisture, pH, and electrical conductivity (EC)
  • Generate crop recommendations based on soil conditions
  • Store all data and recommendations locally when offline
  • Synchronize data to the cloud once Internet access is available
  • Visualize data on a Mapbox map via an admin dashboard, showing coverage and crop suitability per region

3. How we built it

3.1. Hardware

  • FireBeetle 2 ESP32-UE: IoT microcontroller (16MB Flash, 2MB PSRAM)
  • RS485 4-in-1 Soil Moisture, Temperature, pH & EC Sensor for Smart Agriculture (IP68, 5-30V)
  • DSD TECH SH-U12 RS485 to TTL converter: Enables sensor ↔ ESP32 communication

3.2. Software Stack

Technology Purpose
ESP32 (Arduino) or Arduino UNO R4 WiFi Hosts local APIs and communicates with the sensor
React.js Frontend for Agent and Admin portals
Tailwind CSS Styling for responsive and clean UI
shadcn/ui Modern React component library
Mapbox Displays geo-located crop data
JSON Server Offline local data storage

3.3. Agent Portal Features (Offline React Dashboard)

1. Test ESP32 or Arduino UNO R4 WiFi Connection

  • Confirms local network communication between the app and hardware

2. Start Soil Measurements

  • Triggers the ESP32 or Arduino UNO R4 WiFi to read sensor data

3. Display Real-time Data

  • Receives data

4. Generate Crop Recommendations

  • Processes soil data and fetches best-fit crops

5. Save Measurements Locally

  • Uses a local backend (JSON Server) to persist data offline

3.4. Admin Portal Features

  • Manage field agents by province and city
  • Display recommended crops by region using Mapbox
  • Show analytics: soil health, agent performance, regional trends

4. Challenges we ran into

  • Handling serial communication between the RS485 sensor and ESP32 or Arduino UNO R4 WiFi
  • Ensuring offline-first reliability of the React app and backend
  • Maintaining data integrity during synchronization with the cloud
  • Geo-tagging without GPS, relying on assigned region metadata
  • Stabilizing sensor readings in varied soil conditions

5. Accomplishments that we're proud of

  • Fully functional offline-first system — no Internet needed in the field
  • Real-time parameter reading and visualization on the ESP32 or Arduino UNO R4 WiFi and React
  • Crop recommendations logic based on custom matching algorithms
  • Seamless Mapbox integration for geo-visualization of data
  • Smooth local → cloud sync flow

6. What we learned

  • How to integrate hardware (ESP32or Arduino UNO R4 WiFi) with frontend frameworks
  • Deepened knowledge of serial communication (RS485 to TTL)
  • Built a deeper appreciation for resilient offline-first architecture
  • Realized the importance of UX in rural-tech tools
  • Practiced effective data modeling for soil and crop relations

7. What's next for SoilTrack

  • Implement real cloud backend (e.g., Firebase, Supabase, or a REST API)
  • Add device provisioning & remote updates
  • Improve crop recommendation model using ML and agronomic datasets
  • Expand to GPS-based geo-tagging
  • Add PDF reports & exportable summaries for agents & admins

Key Advantages

  • Offline-first, Internet optional
  • Localized, accurate crop suggestions
  • Simple, agent-friendly UI
  • Geo-visual insight for policymakers & stakeholders
  • Perfect for rural zones with limited connectivity

Built With

  • arduino-uno-r4-wifi
  • esp32
  • mapbox
  • react
  • rs485/4-in-1/sensor
  • supabase
  • tailwind
Share this project:

Updates