Previously, only mathematical statistics was engaged in this, they began to use machine learning and artificial intelligence, which, as methods of data analysis, added optimization and computer science (that is, computer science, but in a broader sense than it is commonly understood) to mathematical statistics.

Contents

What is Data Science? Directory Blog
Generally speaking, Data Science is a set of disciplines from different directions responsible for analyzing data and finding optimal solutions based on them.

What are the sciences in this area doing?

First, programming, mathematical models and statistics. Also you need to learn Data science from scratch. But not only. It is very important for them to understand what is happening in the subject area (for example, in financial processes, bioinformatics, banking or in a computer game), so that even real questions can be answered: what risks accompany this or that company, what sets of genes correspond to a certain disease how to recognize fraudulent transactions or what human behavior matches the players to be banned.

Why is this necessary at all?

First of all, thanks to the analysis of big data, it is possible to make more efficient decisions. This, for example, was shown by the recent election campaigns in the United States: using algorithms based on data sets, it is possible to capture the mood of the audience and more accurately target advertising messages (which was probably demonstrated by Donald Trump's team during the election campaign).

The benefits of data analysis can be obtained in all more or less applied areas where there is enough data. For example, in medicine, algorithms make it possible to better diagnose diseases and prescribe a treatment plan. Human resource management can be improved if algorithms help to identify in advance that communication problems have begun in the team.

And when did they start using it?

Recently. With the growth of both the volume of data and computing power, it has become possible to more efficiently solve old problems. Many of the algorithms used today have been known for more than a dozen years, they just became more relevant and efficient. Machine learning algorithms require a huge amount of information. Image recognition with more accuracy than humans, more accurate translators and more recent weather forecasts, all look like a space rocket for which the right fuel has finally been found. Read Python workout PDF for example.

But decisions are still made by people, right?

Most of the time now, yes. But in general, with sufficient technical knowledge, it is already possible to automate the adoption of simple decisions - where there are clear, executable rules. For example, cybersecurity systems today run almost entirely on machine learning algorithms, deciding whether to send an email to spam or block a questionable transaction. Of course, they do this based on the data they already have. The next step in using Data Science is to automate more complex decisions or create a smart assistant. This is approximately how navigators work now, but you can still remember T9 on old phones, which learned our phrases and adjusted. This is followed by the automation of task chains or even specific professions.

How is this applied in the banking sector?

First of all, this is the so-called credit scoring, that is, an assessment of the borrower's reliability for making a decision on granting a loan. It takes into account not only education, work experience and average salary, but also several dozen indicative factors. Another important feature is the fight against credit card fraud. The algorithm processes tens of millions of transactions through itself every day and makes decisions on them.

Data Science and, in paricular, The Python Workshop applies to other financial services as well. For example, based on the client's travel history and medical care requests, it is possible to determine the likelihood of an insured event and, as a result, the cost of insurance for him.

What skills do you need to have to work with big data? Everything is as we wrote: you need to know a little programming, be able to analyze data and interpret the results. And there are no problems with practice in Data Science - in the world there are many platforms for online contests with real problems, where you can try your hand at any time: for example, in word processing, bot breeding, or teaching a neural network to recognize cats in photographs.