Description
PyData Amsterdam 2016
Description
In this talk I will share some of my favourite tools and tricks I use every day as Data Scientist. They help me to solve all kind of problems, from statistical modeling all the way to scalability issues. Expect machine learning, math, algortithms and of course python. All of them are necessary to be a Pragmatic Data Scientist.
Abstract
The Pragmatic Data Scientist Catalog:
- kNN: from slow to fast using math.
- Clustering: From k-Means to Spherical Clustering in a breeze.
- Missing values in classification: from bad to good using old truth.
- Power law: Bucketing the beast.
- The King of proportions Ranking.
- Hyperparameter search: one tool to rule them all.
Python code for all tricks and tools will be available in github for everyone to use change and challenge.