This dataset supports our empirical study on how students engage in software modelling, focusing on their actions, decision-making processes, and challenges encountered. The data was collected from students at two universities as they worked on class and sequence diagrams. The dataset includes:1. A complete log of modelling actions(Time - Action Log_Exploring Actions, Interactions and Challenges in Software Modelling Tasks: An Empirical Investigation with Students) (Add, Delete, Edit), along with timestamps.2. An excel sheet (refer to probelm_diagram)containing the pairs, two codes for the challnege "interpreting modeling problem": refer_problem, refer_diagram, and supporting statements from student pairs.3. An R script (Co-occurance_CD_RU-pair 1) to produce co-occurancy table for one student pair from RU, for class diagram. The script is reproducable and access data from the Time - Action Log_Exploring Actions, Interactions and Challenges in Software Modelling Tasks: An Empirical Investigation with Students. The script produces cooccurancy table for each sheet, as a co-occurrence matrix of modeling actions during the creation of each diagram.By providing this dataset, we aim to enhance transparency and replicability in qualitative research on software modelling. This dataset complements the findings presented in our manuscript by offering detailed empirical evidence to support our conclusions