This project is a showcase of how Edge Impulse can be used along with the Arduino Edge Control board to monitor the current environmental conditions and make inferences about when water is needed. The device is also fully off-grid as it uses a battery and solar panel for power. You can download the machine learning model as an Arduino library here.

• Arduino Edge Control
• Arduino MKR WiFi 1010
• DHT22 temperature/humidity sensor
• 12V sealed lead-acid battery
• 18V solar panel
• Hydrostatic watermark sensor
• Outdoor junction box

• The edge-control folder contains the program that runs on the Arduino Edge Control. It requires the Arduino_EdgeControl library that can be found in the Library Manager. The hydrostatic watermark sensor gets connected to CHAN01.
• The wifi-board folder contains the program that runs on the Arduino MKR WiFi 1010. It uses the onboard WiFi module to communicate with the server found here. The board also gets connected to a DHT22 sensor and it sits in the MKR1 slot on the Edge Control. Make sure to replace your WiFi credentials in arduino_secrets.h and the server's IP address on lines 25 and 69 within the ino file.
• gardenedge-server contains the app which is run with $ python3 app.py. It listens on port 8080. Each time the program is run, a new CSV file gets created and added to as more data comes in from the device.
• Finally, data-visualizer is a program that allows you to take your datapoints and see them on a graph plotted over time. Run it with $ python3 visualizer.py DATAPOINTS_CSV_FILE where DATAPOINTS_CSV_FILE is the path to your datapoints file that the server produced.