Cyberbully Plant

Overview

This project leverages IoT sensors, cloud-based data storage, and automated social media interactions to maintain consistent plant care without diminishing user involvement. A Raspberry Pi Zero, working in concert with soil moisture, light, and temperature/humidity sensors, gathers environmental readings at regular intervals and uploads them to MongoDB Atlas, where they are accessible for real-time analysis and display via a user-friendly web interface. Whenever the system detects that the moving average of sensor data falls outside user-defined thresholds, it invokes a GPT-based script that generates witty, passive-aggressive tweets, publicly nudging the owner to take corrective action. By complementing automated data collection with humorous social accountability, the system balances technological convenience with human engagement rather than eliminating the personal touch inherent in nurturing a plant.

Key Features

Multi-Sensor Setup

In addition to using a capacitive soil moisture sensor to track hydration needs, a TEMT6000 module to measure ambient light intensity, and a DHT22 unit for temperature and humidity readings, the Raspberry Pi Zero polls these sensors at set intervals, collecting analog and digital signals (via an external ADC for analog measurements).

Cloud Storage & Averages

All sensor readings are packaged into JSON and sent to MongoDB Atlas, where they are stored and used to calculate daily moving averages. This filtering smooths short-term fluctuations, preventing false alarms when conditions briefly dip or spike.

User-Defined Thresholds

A dedicated web interface allows the user to establish or modify “ideal” ranges for moisture, light, and temperature. The system automatically triggers GPT-based tweets if readings deviate persistently, ensuring that minor changes do not lead to excessive alerts.

GPT-Driven Tweets

Should the system detect sustained suboptimal conditions, it contacts GPT to generate humorous, passive-aggressive tweets that are then posted to Twitter. By harnessing social pressure in a lighthearted way, the project motivates prompt attention to the plant’s needs, providing an entertaining alternative to fully automated watering systems.

Design Process

Objectives and Requirements

During the early stages, the objective was defined as merging the advantages of IoT automation with genuine user participation, focusing on data-driven thresholds and public notifications to highlight the plant’s condition rather than silently resolving issues.

Hardware Selection & Calibration

A Raspberry Pi Zero was chosen for its compact form and Wi-Fi compatibility. Soil moisture, light, and temperature sensors were tested to ensure reasonable accuracy, with an external ADC bridging analog signals for soil moisture and light sensing.

Cloud Integration & Analysis

A MongoDB Atlas cluster was set up for time-series data storage, receiving sensor information via HTTP or MQTT. A backend service then computes daily moving averages, checks thresholds, and invokes GPT whenever conditions remain suboptimal for a sustained period.

Social Media Posting

The GPT service produces tweet text, which is posted to Twitter through secure API credentials. This ensures user engagement is maintained without risking private data exposure. The posting frequency and tweet style were refined based on early testing feedback.

Web Application & Testing

A user-friendly web interface displays environmental trends, allows on-the-fly threshold adjustments, and logs triggered tweets for transparency. Pilot tests with a single plant guided calibration of sensor readings, tweet frequency, and threshold logic. Encryption safeguards data transit, and environment-based credential storage protects API keys.

Security & Privacy

Data is sent securely using encryption, while Twitter and GPT API keys remain stored in protected environments. Users can manage tweet settings to avoid over-sharing or to disable the public accountability feature if desired, striking a balance between convenience, control, and playful social nudges.

By coordinating sensor-driven data collection, dynamic threshold analysis, and socially oriented reminders, this project ensures that plant care remains a human endeavor enhanced by IoT capabilities rather than overshadowed by them.