comments

By Gregory Piatetsky.

I recently came across a good and brief explanation on Quora by Mateusz Opala - "How to become a Data Scientist" and we share it here, with added links.

Briefly, the most important steps:

• learn Python, it's great for preprocessing, has many machine learning/image processing/nlp libraries. Unless your data is really big, you can easily use Python. Strongly suggest for prototyping. Get know libraries like scikit-learn, scikit-image, Theano. Try OpenCV in Python.
• Machine Learning is the most promising part of AI currently. You should get deep understanding of it, and it will take a lot of time. I would recommend Andrew Ng's course from Coursera, then Geoff Hinton's course also from Coursera.
• You can't become machine learning expert without learning Bayesian reasoning. For beginners: Bayesian Reasoning and Machine Learning: David Barber. If you already have deep math (statistics, linear algebra) background you should try Kevin Murphy's Machine Learning: A Probabilistic Perspective. There's almost everything.
• Get your hands dirty. Implement most important methods by yourself to get more understanding, experience and fun :>. It's really great memory when I saw first time that my neural net can recognize digits.
• Try to be up-to-date, read papers, Quora and follow great data scientists on Twitter.
• There's Udacity nano degree for data analyst

Quora original.

• 7 Steps for Learning Data Mining and Data Science
• Will Deep Learning take over Machine Learning, make other algorithms obsolete?
• Most Influential Data Scientists on Twitter and Quora