Used Scikit Learn and Pandas to predict the light status (ON/OFF) of a user in his/her home using different machine learning algorithms
Created a dataset of a user with his light usage for a week (granularity of every 10 seconds).
Decision Trees (100%) performed better than Linear SVC(97%)
Would have been better if had more data
Communication between 3 languages
Nest API not super friendly when writing values
Getting different parts to work together
Setting up the networking for the project
Making different hardware work with Android App.
It's a protocol providing full-duplex
communication channels over a single TCP communication. Standardized in 2011
Initial purpose to be implemented in web browsers and web servers but now also used in client and server applications.
Real time backend database for Nest Cloud data access.
Data syncronization across devices is easy and lends to quickly building applications.
Nest uses OAuth 2.0 authorization for the thermostat.
Python Flask Server/Client on Intel Edison
Arduino C Server and Light Control
Android App to control the lights and use Nest API
Socket.IO libraries for Flask and Android
Make a work with Nest API product
Pandas and Scikit learn for Machine Learning
Intel Edison - Atom Dual Core Processor,
1 GB DDR3 RAM, Bluetooth 4.0, WiFi, Yocto Linux OS
Nest Thermostat - Google Nest Smart Thermostat, App Controlled, Machine Learning to predict target temperature.
TILE Arduino Lights - Arduino connected smart lights with different colors.
Estimated 50 billion devices on the network by 2020.
Network of Physical objects embedded with electronics and software which enables them to collect and exchange data.
Very diverse application area, like farming, waste management, smart home, smart cars, etc.
What components make up the project?
How all components come together to
make a system which solves a problem