The diversity of students’ learning profiles and the limited use of data-driven diagnostic assessment in mathematics classrooms present significant challenges in implementing differentiated instruction. Current classroom practices are still insufficiently supported by digital systems capable of accurately identifying students’ learning needs in real time. This study aims to develop SIDIC (Student Identification for Diagnostic Instructional Classification), a learning analytics–based digital diagnostic assessment system designed to identify junior high school students’ mathematics learning profiles and support data-driven differentiated instruction. The study employed a research and development approach using the ADDIE model, consisting of analysis, design, development, implementation, and evaluation stages. Instrument validity was evaluated by seven experts across 40 items covering 10 aspects, resulting in an average Aiken’s V of 0.85, indicating high validity. Practicality testing involved 92 students and 9 mathematics teachers from three schools, showing that SIDIC falls into the very practical category (student score = 4.33; teacher score = 4.49). The findings indicate that SIDIC is easy to use, efficient, and effective in identifying students’ learning profiles. Overall, SIDIC represents a valid, practical, and adaptive digital assessment innovation that bridges diagnostic assessment, digital technology, and data-driven differentiated instruction within an integrated system.