15 steps to become a Data Scientist

Data Scientist, Machine Learning, Deep Learning, A.I? ....

Have you heard these words lately? I hope you do because within the next year or two it will become the HOTTEST job/career on the computer science/developer world.

I am introducing in this blog: 15 steps to begin a career as a "Data Scientist". I said "begin" because, just as in any other field or industry, it is the experience and 10,000 hours of practice (read Malcom Gladwel's "Outliers") what will make you GOOD at something.

The world wide web is infinite, and sources of information about "data science" are uncountable. Some sources better than others obviously, but regardless the amount of information is A LOT. Is for this reason that I am writing this blog post. After many months of reading and researching, I put together below what I consider a decent study plan that will give you an excellent foundation to begin the journey in the fabulous world of Data Science and Big Data.

Before starting on this plan, let me make something very clear and be very blunt about it: This is not for anyone. The first thing you need to have before even thinking about being a data scientist is a college degree in Computer Science, Engineering, Math or Statistics.

If you don't have or plan to have any of these degrees, I sincerely, from the bottom of my heart, recommend that you do something else :)

Ok, so, without more introduction, here you go: 15 Steps to become a Data Scientist:

1) Read this book: Data Science from Scratch

2) Complete, with the certificate, the following online courses:

- Data Science Machine Learning Essentials

- Introduction to Data Science

- Intro to Machine Learning

- Neural Networks for Machine Learning

- Deep Learning

- Reinforcement Learning

3) Readings on RNN (Recurrent Neural Networks):

- Understanding LSTM Networks

- A beginners guide to Recurrent Networks and LSTM

- The unreasonable effectiveness of Recurrent Neural Networks

- Recurrent Neural Network Tutorial (Part 1,2,3,4)

4) Study and Practice on these Libraries and Machine learning tools:

- H2O.ai (Deep Learning, GBM, GLM and Advanced algos)

- TensorFlow

- Keras (Deep Learning) and Elephas: Keras on Spark - Lasagne (Deep Learning)

- XGBoost (Gradient Boosting)

- SciKit-Learn

- Azure ML

- Spark MLLib

- BURLAP- Reinforcement Learning

- PyBrain – Python libraries for Reinforcement learning

- ConvNetJS

- DL4J – Deep Learning for Java

5) Complete the following training on Spark:

- Spark Summit Part 1 - Training

- Spark Summit Part 2 - Training

- Read this book: Learning Spark

6) Learn about DataBricks (managed Spark Clusters):

- Spark Summit 2015 demo

- Spark Summit Europe 2015 demo

- Spark Summit East 2016 demo

- Complete this online course on EDX: Distributed Machine Learning with Spark

- Open a "Comunity Edition" acccount on Databricks and complete all the examples and training.

7) Training on H2O.ai:

- Scalable Data Science and Deep Learning with H2O

- H2O with R Demo

- Building a Smart Application with H2O

8) Databricks + Spark + H2O.ai:

- H2O Rains with Databricks Cloud for Spark

- H2O over spark (using Databricks and AWS)

- Examples to run on Sparkling water

9) Training on Keras and Lasagne over Theano:

- Deep Learning with Keras

- ML going deeper with Theano – PyCon

- Theano Tutorial

- Lasagne Tutorial

- Examples – Notebooks on NNs with Theano and Lasagne

- Setup a Deep Learning environment on Windows (Theano and Keras with GPU enabled)

10) Training on TensorFlow:

- Oficial Tutorial

- Scikit Flow – simplified interface for TensorFLow

- Installing TensorFlow with Python 3 on EC2 GPU instances

- Examples in Github

- More good examples in Github

- Deep Learning with Spark and TensorFlow

- Machine Learning with Financial time series Data

11) Training on Data munging, formats and Visualization:

- Munging, Modeling and Pipelines using H2O

- Data Exploration with SparkSQL demo

- Parquet vs Avro: Format face-off!!

- How-to: Convert Text to Parquet in Spark to boost Performance

- Overview of python visualization tools

- Kaggle Script – Python data visualization on the Iris Dataset

- Kaggle Script – Predictions in the Republican Primary

12) Tips and Tricks:

- Top 10 Deep Learning Tips and Tricks

- Top 10 Data Science Pitfalls

- 6 Tricks learned from OTTO Kaggle Challenge

- Fine tuning a Neural Network

13) Print and hang these images on your wall:

- Image – Machine Learning Algorithms

- Image – SciKit-Learn Cheat Sheet

- Image - Azure ML Algorithm Cheat Sheet

- Image – Road to Data Scientist

- Image – Machine Learning Process

- Image – Azure ML Studio diagram

- Image – Learning Paradigms and Algorithms

- Image – Azure Data Streaming architecture and workflow

- Image – Data Visualization in Python Cheat Sheet

14) Training on Kaggle:

- Machine Learning Best practices we’ve learned from hundreds of competitions

- Kaggle FAQ

- Data Exploration with Kaggle Scripts

15) Compete on Kaggle and Cortana Intelligence competitions until you reach top 10