How to predict your location with wifi signals

El machine learning every day it advances more, a couple of days ago I met an application called whereami, which learns the locations that we indicate, for this it uses Wi-Fi signals and artificial intelligence algorithms.

What is Whereami?

whereami is an open source, cross-platform application (GNU / Linux, OSX and Windows), written in Python by Pascal vanKooten, which uses Wifi signals and machine learning to predict where you are.

Thanks to the class sklearn's RandomForest, this application can predict up to very close locations (from 2 to 10 meters), that is, whereami it can tell if you are on your bed or at your bedroom table desk.

This application can be integrated with others and can be the beginning of a myriad of new applications or functionalities.

Whereami uses

An application that allows us to predict our location can have various uses, among which we can highlight:

• Show off with our friends that our computer knows at all times where it is in the home.
• It can have a wide use in roboticsFor example, the robots that are in charge of cleaning can be programmed so that depending on their location, they carry out the cleaning or not.
• Home automation can be favored, since depending on our location functionalities of our home can be activated. (Imagine sitting in your armchair and automatically turning on the TV, turning off the lights, activating the air conditioning, cooling the beers and stopping counting).
• We know that GPS already exist, but for example in the event of a catastrophe, knowing in which specific area of ​​a hotel you are located would be an incredible help in saving lives.
• The limit is your imagination.

How to install Whereami

Installing Whereami is extremely easy, just have python installed in your distribution and run the following command:

pip install whereami

How to use Whereami

For whereami go learning or show us our location we can use the following commands:

# Take 100 samples from her bedroom
whereami learn -l bedroom -n 100

# Take 100 samples from your kitchen
whereami learn -l kitchen -n 100

# validation with the data
whereami crossval
# 0.99319

# If you want the computer to say where it is located
whereami predict | say
# Computer says: "bedroom"

whereami predict_test
# {"bedroom": 0.99, "kitchen": 0.01}

If you want to delete the information learned by the application, you can check the following folder $USER/.whereami.

I hope that from now on your computer will learn each of the places where you use it on a daily basis. What use would you give this application?