How to become a data scientist and what is this profession?

Big difference between “data science” and statistics no, but we have to captivate young people new-fangled terms, and now is Analytics-statistics appears in the job “data scientist”.

Its task is to plunge into the bottomless ocean of customer data and find the insights to increase the profits of the company. It seems about the same spoke of the “Grose-hackers” — literally “growth hacking”, that is, those who know how to find some sort of super-idea that increases sales by hundreds of times. To Grose-hackers engaged managers and Directors on development, and then they found a new name. Задача специалиста - погрузиться в огромный объем данных и взять нужноеA specialist task is to take in a huge amount of data and take the right
However, the “Grose-hackers” do not know how to delve into statistics, and more generally in the Humanities. Visual difference between the specialties is visible visually: “Grose-hacker” creates many beautiful presentations with points of growth, data scientist draws abstract pictures of the points with comments like “I didn’t use the p-value, since they can result in false positive”.

Another date-scientist is able to build hypotheses and ask questions. A specialist collects a large number of unmanaged and unstructured data, converts them into an understandable uneducated masses format to solve specific business objectives of the customer research.

An employee applying for a position in the field of Data Science, must know the programming languages. Preferred R and/or Python, plus query language (come even good old SQL).

Next you need to test the candidate’s knowledge in statistics (after all, data science is statistics). Methods of testing hypotheses, normal distribution, types of group sets and other mathematical wizardry.

Because it will work with big data (BigData), then treatment methods will be tied to it. Accordingly, the data scientist needs to understand machine learning, neural networks, deep learning. Knowledge meme about Herman Gref (the”Block chain, big Data, Machine learning, Deep Machine learning, Agile”) is a necessary but not sufficient condition!

Mathematical modeling, construction of hypotheses, analysis of the sample, the provider, method of least squares, simplex method, systems of differential equations of the second order — all this instrumental some clever way to help the company avoid risks and to destroy competitors.

Математическое моделирование, построение гипотез, анализ выборки и прочие методы помогут сделать правильный выводMathematical modeling, construction of hypotheses, analysis of sampling and other methods will help you make the right conclusion

Rumor has it that 10-20 years from each effective Manager will require basic knowledge of Data Science. Now in the West there is a growing demand for managers with knowledge in this area.

However, it is suspected that such statistics create date-scientisty to justify their salaries (rather big). Because every statistic is a kind of lie, to believe in such predictions is not necessary.

But do not rush to hire date-sientist to work! It turns out that the next stage of Data Science is total automation in making difficult decisions. That is soon cientistas will replace the algorithms and working with data will be worth a penny.

Leave a Reply