Fig 1.
Orange workflow for Kaplan-Meier survival curve estimation.
The workflow automates the Kaplan-Meier survival analysis using trainees’ prepared dental filling data. (a) Initially, data created from timeline diagrams is loaded into the workflow. (b) The Kaplan-Meier plot widget calculates and plots the survival curves, allowing trainees to visualize and confirm results obtained from manual analysis. Creating this workflow supports learners in developing an intuitive understanding of how censoring and time-to-event data are translated into stepwise survival curves.
Fig 2.
Orange workflow demonstrating stratification and comparison of survival curves.
This workflow guides trainees through the grouping of data based on a continuous variable. The Select Rows widget partition data into groups (e.g., less than vs. greater than or equal to 6 minutes daily). Kaplan-Meier plot widget visualize survival curves and computes the log-rank test. Trainees can interactively select different variables to explore their impact on patient survival, and by using the workflow understand how feature-based stratification influences survival curves and statistical significance.
Fig 3.
Gene set enrichment analysis workflow for survival prediction.
The workflow guides trainees through the process of calculating ssGSEA enrichment scores for a specific gene set and stratifying patients based on these scores. The workflow stratifies patients based on ssGSEA scores and visualizes Kaplan-Meier survival curves, revealing the prognostic impact of biological pathways. Users can experiment interactively with different gene sets to identify biologically meaningful predictors of survival. We use this workflow to support learners in understanding how pathway-level features may provide more robust and interpretable survival predictors than individual genes.
Fig 4.
(a) Overview of the user study: a pre-test questionnaire, four content chapters, and a post-test.
The tutorial includes approximately 40 min of video and 22 exercises, progressing from basic survival analysis to biomarker discovery. (b) Distribution of time spent and scores: participants spent an average of 88 min. With a 22-point maximum, 93% (113 participants) passed by scoring above 17.