Taking care of plants can be a great headache, especially if you have a variety of plants in your big backyard.
What it does
Our hack consolidates data over a week from the IBM weather Insights Service and data from water sensors attached to sprinklers (for when the last time it was turned on) and reaches a decision for when the backyard should be watered. Additionally, we also provide curated service for how to take care of the different plants in your backyard by using a chatbot interface to allow the user to input everytime a new plant is added to their backyard. The service takes into consideration the temperature, humidity, soil moisture and season as parameters from the weather api which is then turned over to looking at the different planst and the decision matrix of the care to be taken for each plant is based on the plant age, if it was given plant food recently, amount of sunlight it has recieved over the span of the last month, if it needs to be trimmed and if it needs to be reloacated depending on the season.
How we built it
This is a service oriented architecture. We intend to build this using data from the external input interfaces (our raspberry pi connected to the water sensor using the Jillia to push the data off to IBM Watson). IBM Watson, runs (concurrently) the service of the chatbot that polls for an event of the user/owner to input the addition of a new plant to the backyard.
Challenges we ran into
Setting up and the learning curve needed for IBM Bluemix. Getting valid data from the sensors.
Accomplishments that we're proud of
The system is a fairly complex one once we drilled down on the requirements. Hence we are proud of the decision flow introduced.
What we learned
This learned lots about the protocols for communication between hardware and the cloud.
What's next for Plant Care
We expect this tool to hold further applicability for large-scale commercial operation, in which data on outcomes can used to train more robust machine learning models.