FsLab: Doing data science with F#

Опубликовано 27 июня 2016, 20:15

How to get knowledge from data? We need to access CSV files and REST services, combine the data while handling missing values, try different analyses and machine learning algorithms and then build visualizations to make our point. We need to explore data interactively, but end up with reproducible scripts that can be easily deployed in production. I’ll demonstrate end-to-end data analysis using FsLab – a cross-platform set of data science libraries and tools based on F# that make it easy to perform the entire process with a single tool. Type providers turn external data sources into inherent part of your language; integration with tools like R gives you immediate access to professional packages and HTML5-based visualization tools produce beautiful results. Along the way, we’ll explore correlations between countries using the WorldBank, we’ll look at survival rate of different passengers on Titanic and we’ll look how different political parties contribute to country’s debt.