Derived KPIs to identify at-risk students using data collected from a specially curated C++-based coding competition using attributes such as students self-reported pre-test and post-test confidence + telemetry (no. of compiles, scoring efficiency, time taken) using Python.