how to become a data scientist

I have recently started competiting on Kaggle. I have been doing data science for 20 years and have a masters in data science. I still learn a lot from competiting. What is great is that kaggle is available no matter your experience level. If you are just starting data science, Kaggle is a great way to learn about data science. I would recommend start working through their tutorials and then enter a competition. Read the discussions. Download the code that others use and work through it on your own to make sure you understand.

Hi Moulay,

Great question. In general Kaggle competitions provides a great introduction to Data Science , is a fantastic place to hone your modelling skills and build a portfolio. While some folks have had success breaking into the data science field, it's not a guaranteed way to get a job. Professional data scientist usually develop other skills which are not prioritized and learned in Kaggle competitions. Some of these skills include wrangling and cleaning complex data, SQL for database querying and analysis, general software engineering skills for writing testable and modular code and deploying models in production settings, and work with distributed systems for large data sets.

That being said, you should defintinately invest in learning more about data science through Kaggle competitions. But to make yourself competitive for Data Science jobs, also focus on picking on data clean and wrangling skills (using a language like Python or R), basic algorithms and data structures knowledge found in standard computer science curricula, SQL, and general good coding skills (using git for code maintenance, writing good unit tests, modular and testable code, etc).

Good luck !

A service like Kaggle is useful in becoming familiar with Data Science methodologies and techniques, but it is not the only thing you should be doing to show an interest an aptitude for the discipline.

You will need to show recruiters that you have a broad capability in the form of mathematical knowledge, programming skills - a personal github account is a good way to showcase your experiments and investigations, and experience with data processing and manipulation tools such as databases and SQL.

Publicly available datasets are a great source of material for your experiments: https://github.com/awesomedata/awesome-public-datasets