Abstract: We will compare some of the most popular R Machine Learning meta-libraries in a workshop with live coding where the audience is invited to code along.
Our goal will be to create a smartphone-based positioning system using a publicly available Wi-Fi fingerprinting dataset (UJIIndoorLoc database).
In the first part of the workshop we will focus on understanding, cleaning, visualising and transforming the data (including dimensionality reduction).
Once the data is ready, we will train classification models (to predict building and floor) and regression models (to predict latitude and longitude) using different R frameworks. We will evaluate our models and make a prediction that can be sent to the researchers of Jaume I University who created the dataset to compete in their ML challenge (they keep a private validation dataset).
Bio: In the spring of 2016, an enlightening work experience at a London-based Data Consultancy firm pushed Guillem towards a big career leap: from Translating and Social Science Teaching to Data Science. Three years later and having worked as an analytics consultant at Accenture, Guillem is the Senior Data Analytics Mentor at Ubiqum Berlin.