Understanding job requirements is essential for establishing and optimizing employability-oriented education programs. Most relevant research focus on clarifying the skill requirement of an occupational field. In this research, we argue that job tasks serve as a bridge between a job and the required skills, and we provide a method for investigating the job-task-skill pattern of job requirements using text mining on publicly available job advertisements. To provide this, we: 1) collect data on thousands of job advertisements through web crawling and scraping; 2) categorize the jobs through title analysis; 3) identify the topic of both tasks and skills through word co-occurrence network clustering; and 4) systematically analyze the characteristics and internal relationships between job roles, tasks, and the required skills. A test case was conducted in the context of China's big data sector, and the findings show that the proposed strategy is viable, practical, and instructive.