The expansion of AI and Data Science has led to a higher demand for data science related professionals. This surge has created a need to identify the knowledge, skills, and tools required for these roles. This paper proposes a data-driven approach that employs classification models to analyze job postings and extract key competencies. The analysis is based on data from the Polish job market. The methodology involves feature selection using a large language model to extract keywords representing skills, tools, and knowledge from job advertisements. A neural network classifier is then trained to predict job titles based on these features. Feature importance analysis, utilizing permutation importance, is conducted to identify the most influential factors in the classification. The findings from weight magnitude analysis and occlusion sensitivity analysis provide insights into the distinct characteristics of Data Analyst, Data Engineer, and Data Scientist roles as reflected in job postings. The analysis reveals the core competency requirements for each job position, contributing to a clearer understanding of the evolving demands within the data science field.