<p>With the quick development of the digital economy, software project managers take a pivotally important position in guaranteeing project success, yet traditional competency models for this role have shortages like dependent on subjective judgments and insufficient quantitative support. Therefore, this study sets its aim to build a scientific and practical competency model for software project managers by combining big data technology to solve the shortcomings of conventional methods. To reach this goal, a two-stage research design was utilized. In the first stage, web crawling technology was used to collect large-scale recruitment data from online recruitment platforms, and text mining based on Python was applied to analyze unstructured information such as job descriptions and requirements; hence, key competency indicators were extracted. In the second stage, questionnaire surveys were sent to software project managers and senior engineers to gather subjective evaluation data. Then, SPSS software was used for data analysis: factor analysis was adopted to decide objective weights, the analytic hierarchy process (AHP) for subjective weights, and K-means clustering was used to provide exploratory evidence on the internal grouping plausibility of the competency model, with final weights obtained through a combination of subjective and objective methods. Thus, the results show that a competency model for DW Company’s software project managers was constructed, which was informed by competency elements extracted from a large-scale industry-wide recruitment dataset. This framework may provide useful reference for talent management in similar software enterprises, while broader generalizability needs further multi-company testing.</p>