At the beginning of 2021, there was a clear imbalance between supply and demand in the labor market in the IT industry. Developers, data analysts and internet marketers have become much more in demand as a result of the pandemic, when businesses began to massively migrate online. We talk about how the market has changed recently and what companies are doing to retain and hire the best people.
Data Scientist is one of the fastest growing specialties of the 21st century. Frost & Sullivan predicts that the big data analytics market will grow by an average of 35.9% per year over the next 10 years.
In this article, we will look at how much money a data scientist can get (spoiler alert: a lot), what requirements are most often in vacancies, how to come to DS and where to develop. Ready? Go!
A data scientist is engaged in the analysis of data arrays – Big Data. Using mathematical analysis and identifying patterns in the data, the data scientist creates models to solve specific business problems or problems.
In fact, the profession of data scientist is closely related to data analyst (Data Analyst or DA) and data engineer (Data Engineer or DE). So closely spaced that most grocery companies don’t separate them at all. And the Data Scientist often takes over the DA and DE responsibilities.
In short, this serious person perfectly understands the data, takes the maximum of useful information from it and knows exactly how to turn it into profit.
Do not regard the position as technical. The main role of such specialists in the qualitative modernization of the business is a detailed strategy and its effective implementation.
Let me explain with a real example, which I saw in an interview with Stephen Brobst, Chief Technology Officer of Teradata.
Imagine a telecommunications company where subscription fees are the main source of income. On the 5G approach, some new network infrastructure, which requires an impressive financial investment. To increase profits, you need to attract as many customers as possible, while simultaneously reducing the cost of services.
The goal of CDO is to profit from a huge amount of data while maintaining its privacy and security.
Information on subscribers is in front of his eyes, he knows everything about them: location, average call duration, tariffs used, clicks on promotions, and more.
It turns out that CDO is changing the traditional business culture, creating new products and services, while improving existing ones through a strategy supported by inferences from data.
CDO is not about education, but about personality. The person who is leading the digital transformation in the company is always ready for challenges. This is a bright, charismatic leader. He knows how to communicate, persuade, overcome resistance, make quick decisions and quickly adapt to changes. Most often, the start begins with the CIO and its division – such people feel trends at their fingertips and are accustomed to constant learning, which is why it seems logical to transform the role of CIOs into CDOs.
Another good candidate type is those with financial experience. After all, the most common reasons for the transformation of companies are the desire to speed up and simplify processes, increase profitability, and improve the quality of services or products. A person with financial experience and education will be able to build working models for such purposes. Schneider Electric’s director of digital transformation began his career in the finance department.
The popular American career site Payscale lists the average CDO salary in the United States – $ 177 thousand per year. For a month, such a specialist will receive about a million rubles at the rate at the beginning of 2019. And, as you can see, this is not even the limit.
Machine learning is an important part of the data scientist profession. Neural networks are becoming more and more popular for analyzing data sets, so a specialist must be able to work with them.
One of the main goals is to achieve a business result. After all, it is with the help of Data Science that predictive models are developed. For example, the behavior of users on the network, exchange rates, stock prices and much, much more. It is data scientists who have developed YouTube recommendation algorithms and improve Google’s search results.
We analyzed over 400 vacancies for a data scientist position opened in October-November 2020.
A clear boundary between the specialties of data scientist, data analyst and data engineer exists only for IT companies and large corporations with large IT departments. Therefore, job vacancies for data scientists often come across tasks that are more appropriate for an analyst or data engineer.
First, let’s deal with hard skills.
High math skills. Higher mathematics, probability theory, mathematical and applied statistics are a must-have for a Big Data analyst. Over 60% of vacancies directly indicate the need for good mathematical training or require a bachelor’s or master’s degree from a university in mathematics, engineering or information technology.
Python and libraries for data analysis and machine learning. Python is listed in 81% of jobs. Also, most often employers require knowledge of special libraries: TensorFlow, Keras, PyTorch, LightGBM, NumPy, SciPy, Pandas, sklearn.
In about a third of vacancies, employers indicate knowledge of Python and / or R. But specifically R (without Python) is rarely requested – only 12 vacancies out of 400. Other programming languages are required in about 3% of cases.
SQL. Databases are the backbone of DS. Therefore, skills in working with relational databases are needed in more than 73% of vacancies. NoSQL databases are less popular – they are needed less than 10% of the time.
Excel stands alone. Although it is not included in the stack of required skills, some companies build data analytics in it. Why this is so is not clear. Perhaps they are simply confusing the functions of data analyst and data scientist.
Data visualization systems. As a data scientist, it is important not only to create working models and forecasts, but also to be able to present this to management. It is desirable in a clear and simple way. Most companies (55%) simply indicate “data visualization systems” – for them it is absolutely not important which ones the applicant will own. But among the most popular are only three – Tableau, Metabase and Power BI.
Machine learning and Deep learning. Machine learning and deep learning are important. Almost 40% of companies separately emphasize that the applicant needs to understand at least in general terms how it all works and how to use it in business.
Many companies point out the need for knowledge of technical English with a level not lower than Intermediate. Conversational skills are often not needed, but technical documentation will have to be read. Moreover, almost all new developments in Data Science are published in English, so you need to understand it at least at an intermediate level.
For hard skills in general, that’s all. Well, or almost everything, because other options are found in less than 10% of vacancies.
In general, they are quite expected. Here are the most common ones that companies ask for: