Ever since the development of the smart thermostat concept a few years ago, several platforms have been working to integrate control of these thermostats on their devices, such as Amazon with their Echo lineup or Google with their Google Homes. Although the voice control industry of these smart thermostats seems to be saturated, we discovered that the text-enabled industry wasn’t. Therefore, we decided to develop a virtual chatbot, able to run on several platforms, that would help users control the Nest Smart Thermostat.
What it does
Our Nest Chatbot allows anyone to converse with a bot, and in the process, be able to control a myriad of processes around the house. The Nest Bot is able to understand the basic commands of controlling a Nest Thermostat, from heating up a house to telling you the current temperature. Additionally, the Bot is accessible from a number of different applications, including Skype, Twilio, and Facebook.
How we built it
We began the development with Microsoft’s bot tutorial. From there, we learned about the basic features and framework of bots. Following this, we continued to build features of our bot that would be specifically targeted towards the Nest. Using the Nest API, we were able to connect our bot’s understanding of language (i.e. Microsoft LUIS) with changes in a physical Nest Thermostat.
Challenges we ran into
As a group, all of us were quite new to the concept of bots and 3rd party APIs. Delving into completely new concepts, we were met with a countless number of errors that often had us stumped. Thankfully, there were mentors around that were able to help us comprehend where we took a wrong turn and lead us back on the right track. Additionally, not having an actual Nest Thermostat hardware device on-site made it difficult to test our code. For the first half of the hackathon, we connected directly to one of the group member's Nest devices at home and received feedback from their parents. Later, when we discovered it, we began using a Nest Virtual Simulator to help us analyze thermostat changes.
Accomplishments that we're proud of
We are quite proud of the fact that we were able to create our very first automatic responding bot. It was quite exciting being able to utilize the Microsoft Cognitive Services to train our bot’s understanding of natural language. Additionally, being able to physically control a Nest at one of our member’s home through a chat system was quite spectacular.
What we learned
Through this entire process, we were able to learn quite a bit about the bot framework and REST APIs. We learned the process by which language is naturally processed and the history of bots in generally. Furthermore, we were able to access data about the Nest Thermostat using the chatbot.
What's next for Nest Thermostat Virtual Chatbot
Nest holds a large variety of products, including indoor and outdoor live cams, that we hope to integrate control of in our future chatbot. The Nest API will allow us to add control of these products fairly simply. We also plan to integrate this bot into Slack.