The data sector, and in particular big data, has been booming for several years. For nearly 10 years, Big Data has been taking up more and more space in companies, particularly due to the explosion in the amount of data being emitted, collected and used. This phenomenon is easy to understand when you know that in 2019, 95% of French people will own a smartphone: a tool for generating a phenomenal amount of data.
According to Valuates Reports, the global big data market was valued at $193.14 billion in 2019 and is expected to reach $420.98 billion in 2027. The market share is so large that competition has become very fierce between the various cloud services such as Microsoft, Amazon, Oracle, IBM, etc. Numerous professions have therefore emerged thanks to this sector. This is the case for the professions of Data Analyst, Data Scientist and Machine Learning Engineer. If you are not familiar with these jobs, then you are in the right place: we explain everything to you.
For this article, we interviewed Clément Renault, Data Analyst at Uni-Médias, a press subsidiary of the Crédit Agricole S.A. group, Amine Hadj-Youcef, Data Scientist at the SNCF and Flavien Gelineau, Machine Learning Engineer at Deeplife, a startup specialising in biotech.
By definition, a data analyst processes various product, customer or company data and its performance in order to derive indicators that will then enable the company to make decisions.
"At Uni-médias, I will study the behaviour of Internet users and then draw up strategic guidelines. This data analysis is specific to the press sector, but we analyse different data depending on the sector in which we work. Generally speaking, I come in after the data scientist," says Clément.
To be a Data Analyst, you almost don't need technical skills. Indeed, it is possible to learn data analysis by practising "on the job". However, it is necessary to know how to master data processing tools (Microsoft Excel) or data visualisation tools (Microsoft Power BI, Tableau, Powerpoint etc.).
Data analysis requires soft skills such as: communication, curiosity, open-mindedness, etc. He adds that there are a lot of things that you learn in the company in this profession and that are specific to a sector of activity.
"I learned my job on the job. I came out of a "classic" web school where I did code, SEO, web marketing, etc. It's only within a company that I'm able to learn a lot. It was only at Uni-Médias that I developed all the skills specific to the needs of the media sector. So I started by capitalising on my soft skills," explains Clément.
Flavien, machine learning engineer adds, "For me, knowing which figure or which graph to choose to illustrate a situation is a real technical skill specific to the data analyst and is of great value. It's taking a complicated situation and pulling out the four numbers that explain that situation while being clear and concise with just those four numbers."
For Clement, the best part of being a Data Analyst is the feeling of being useful which gives him great satisfaction.
"When I have extracted the insights, I am asked to make a recommendation for the company based on the results I have analysed and the competitive intelligence I have done. So at the end of a presentation, my recommendation is discussed and then validated. That's when I feel really useful. The greatest satisfaction I can have is when my N+2, my N+3 or even the General Manager say to me after my presentation "What you said is validated, we believe in it and we're going to do it."
A data scientist combines data from different sources (specific to the business sector and the company's needs) and provides the company with a more global view. He or she then presents the results in a visual way via reports or graphs.
"At the SNCF, there are devices that collect data. As for me, once the data is collected, I will use it to get an overview of how we will carry out certain actions, such as the maintenance of track equipment, for example," explains Amine.
In terms of technical skills for a Data Scientist, Amine explains that it is more than necessary to have notions of programming (mastering languages such as Python, knowing how to install a working environment or a package) but also notions of visualisation and mathematical development (probabilities and statistics in particular).
Clément, Data Analyst, adds: "I think that one of the common points between our jobs is that they all require mathematical skills, particularly in statistics and probability.
However, depending on the type of data being processed, there may be specific methodological skills. This is why we talk about "Data Scientist specialised in such and such a field" and depending on this specialisation we do not expect the same set of skills.
In terms of soft skills, the most important are: communication, curiosity and teamwork. This type of profile never works alone, so it is essential to know how to work in a team.
"Having results is all very well, but you have to know how to communicate them. At the SNCF, I have to communicate my results to the maintenance teams, but this is a public that is not familiar with data science, so I have to communicate in a simple and clear way, with synthetic and easy-to-understand visualisation tools," explains Amine.
According to Amine, the best part of his job depends on the company he works for. In his role as a Data Scientist at the SNCF, the best part of his job is the topics and issues he is asked to think about.
"At the SNCF, I work on future topics such as autonomous trains, optimising the maintenance budget, passenger safety, etc. What I like is that these are projects that are meaningful."
The Machine Learning Engineer designs software that automates predictive models (known as "self-running" software).
To know more: The term "machine learning" comes from the fact that each time the software performs an operation, it reuses the result to perform future operations, increasing its level of precision with each operation.
"I'm going to put machine learning software into production and deploy it so that the whole team can use it," adds Flavien.
A Machine Learning engineer, like a Data Analyst profile, is a job where you learn a lot on the job.
"For me, there's a lot of things that you learn on the job and so you have to love learning, love reading something and understand as quickly as possible 80% of the problem."
Machine Learning engineers can use their expertise in many areas which is why, like a data scientist, they specify the industry sector of the company they work for.
"In this profession, you always have to be integrated into a system. For example, today I am doing cell modelling, so I had to learn and understand biology, at least enough to be able to work in this field," says Flavien.
To be a good engineer in Machine Learning, you have to like learning, like tinkering, like coding and above all have a fairly solid scientific and mathematical background. But there is also a high requirement in terms of speed in this profession.
Flavien explains that his favourite part is the creative aspect of his job.
"Creating a service from scratch, I really enjoy that. I'll take an idea and understand it and then see how I can bring value to it and how to implement it. Once it's implemented, I'm satisfied when I realise that what I've created actually exists and that people are using it and it's serving them."
NB: The jobs of Machine Learning Engineer and Data Scientist are similar in terms of skills, particularly in terms of managing large volumes of data and modelling them. We therefore asked Amine and Flavien about the differences that could exist between their two jobs.
According to Flavien, Machine Learning Engineer, "These jobs are very similar in terms of concept. However, I will be asked to do less modelling than a Data Scientist. Instead, I will be asked to be quick in producing a functional tool. So I'll test an idea, put it into production, run it on different data and cut the project or not."
For Amine, Data Scientist "There is a clear difference between the two jobs. Machine learning is a data analysis technique and the Machine Learning Engineer is specialised in this method. As for the Data Scientist, he is not forced to use only this method of analysis to process data."
At Urban Linker, data-related jobs, and in particular Data Analyst, Data Scientist and Data Engineer, are among the most sought-after jobs for our clients. These jobs are evolving year on year. This can be seen in our annual Salary Surveys: in 2020, a junior Data Scientist earned an average of €46,000 in the Paris region and €35,000 in other regions. To find out more, download our Tech Salary Surveys for the Paris Region.