Data-Driven Organizations: Persistence is Key

Even sparrows on the branch know that the foundation of every modern business, whether manufacturing or service-oriented, is data. There is no business whose products, services, sales and marketing processes, and numerous other activities do not rely on large sets of data that, of course, need to be adequately stored, protected from cyber attacks, and continuously managed. Companies, both small and large, have realized the power that stored data about business processes can have for current operations, as well as for decision-making for the future. This is why data experts, namely data analysts, scientists, engineers, business systems analysts, and database developers have become some of the most sought-after professions in the labor market.

Driven by AI

Current topics such as generative artificial intelligence and ChatGPT have sparked great interest among the general public in the field of analytics and artificial intelligence, but in recent years the market has had significant needs in these areas, emphasizes Andrea Pirša Ilić, head of the Center of Excellence for Artificial Intelligence and Data at A1 Croatia.

– This is best illustrated by the number of students enrolling in data science programs, as well as the opening of numerous studies in this domain. The interest in job openings we create for these areas is truly high, and recently there has been an interesting trend of applications from experienced development engineers and programmers who are deciding to change careers towards becoming data scientists, says Pirša Ilić.

That data analysts are among the most sought-after, if not the most sought-after in the market, is confirmed by Jelena Škalec, analytics technical sales at IBM Croatia.

—

—

– AI is a very current topic – ChatGPT has sparked great interest in AI. Since then, artificial intelligence has been present in conversations around the world, and we can say that it has gained momentum in Croatia as well. An IBM study revealed that CIOs in Central and Eastern Europe see AI, cloud, automation, 5G, and IoT as their biggest investments over the next three years, emphasizes Škalec, noting that many companies she works with already have established data science teams, while many others want to start implementing artificial intelligence into their processes but simply cannot find experts.

An Average Data Day

But what does an average day look like for data experts in different companies, and what knowledge is necessary for someone to embark on this profession? Experts who handle data enhance business operations in organizations reveal this. At Megatrend Business Solutions, which deals with analytical solutions in banking, hospitality, and retail, with an emphasis on big data, AI, and machine learning, Domagoj Marić is employed as the head of the Artificial Intelligence and Data Science Department. He emphasizes that data analysts, by providing quality and timely information based on data from various sources, help clients make decisions, which every director and manager will agree is a crucial part of business. Marić cites the applications Hospitality Insights and ShelfXplore as examples of one of Megatrend’s solutions.

– The ShelfXplore application, based on computer vision, was developed for the needs of distributors and manufacturers of consumer goods, while the Hospitality Insights application helps service industries and companies manage the customer experience of their guests. Data collected through the Hospitality Insights application and from other sources undergoes a cleaning and transformation process and is used to generate interactive dashboards that our clients use to support decision-making at all levels, explains Marić.

For Everyone Involved

Škalec explains that when organizations have prepared data, when they know where each piece of data is located, who has the right to access it, and where that data is used, the processes around creating, for example, reports, data models, or machine learning models are significantly accelerated.

– Research shows that data scientists spend about 80 percent of their time preparing data, and when they have a properly structured data architecture, they themselves spend less time preparing data and can focus more on developing machine learning models. Here we are not only talking about data scientists but about all individuals involved in data usage, such as reporting, because they do not have to wait for someone to prepare the data for them. Therefore, when we have a structured information/data architecture, organizations are more agile and can utilize the data they collect more quickly, explains Škalec, who works on various projects at IBM related to setting up data warehouses, organizing data, and projects in the domain of data governance, noting that recently they have also been working on many projects implementing AI into organizational business processes.

—

—

– With users, we often build the entire story, meaning that with the IBM Cloud Pak for Data platform, we enable the user to connect all end-to-end analytical processes, from the moment the data sits in the database to the moment that data is utilized, adds Škalec.

Nothing Without Analytics

At A1 Croatia, for instance, there is no department that does not use at least some form of data analytics, from various forms of advanced analytics, machine learning model development to the application of artificial intelligence, notes Pirša Ilić. For example, in customer service, they use data science to understand customer needs, regarding the most common needs for contact with A1.

– In email communication with end users, our agents are assisted by internally developed machine learning models that classify emails by content to those agents who best know the domain of the email’s content and who can provide the fastest support to the user. We are also developing internal chatbots; the internal chatbot Bob Rock assists our employees in technical support and various human resources-related questions, while our chatbot Nikša is currently available to users for needs related to vouchers. We have an internal department that deals with advanced analytics, researching machine learning models, and developing solutions based on artificial intelligence. We are developing solutions that personalize offers according to user needs, serve to better understand interactions with users, help optimize sales channels, and manage our network, describes Pirša Ilić.

Domain Knowledge is Key

However, to be successful in their job, every data analyst should possess certain skills. Marić lists several that he considers essential: efficient manipulation and retrieval of data using SQL (Structured Query Language), skills in visualizing data in an understandable way for everyone, problem-solving ability and finding optimal solutions, good knowledge of the business domain of the project they are working on, and ultimately, indispensable communication skills that allow clear transmission of information to all participants in the project. Pirša Ilić also mentions a similar combination of technical, domain, and ‘soft’ skills.

– Distinct analytical skills and knowledge of mathematics are necessary, as well as excellent programming skills and business process modeling. It is very important for a data scientist to understand the data they analyze or model, and domain knowledge of the industry is crucial. Knowledge of cloud technologies, data processing tools, and model development within them also becomes very important skills for data scientists, adds Pirša Ilić, noting that the most challenging part of the job for young data analysts and scientists is precisely the domain knowledge that comes with experience.

Research Spirit

As in almost every job today, data analysts, given the enormous development of technology, must constantly stay updated with the latest achievements in their field. For example, Pirša Ilić follows topics in the domain of analytics and data science on several platforms, conferences in that area, and says that a large part of practical knowledge comes from team members who engage in research.

—

—

– We nurture a research spirit within the department, so that every interesting topic that potentially has application in our development is presented to the rest of the department, and we strive to share some new knowledge. I follow several topics in the domain on LinkedIn, as well as various news through blogs and sites like KDnuggets, Towards Data Science, Data Science Central, Gartner, and similar, explains Pirša Ilić.

In addition to all of the above, Škalec adds that it is very important for data scientists to be persistent.

– Because, for example, data scientists can develop a machine learning model for days, if not weeks, only to realize in the end that it is not good enough, that they are missing some data, explains Škalec.

She emphasizes the importance of thinking outside the box.

– If something has brought value and worked successfully for one user/organization, it does not mean it will work the same way for another. Being a data scientist is not just about programming in Python and replicating similar use cases or copying code; a broader picture is needed, concludes Škalec.