Interactive demo of a weighted linear scoring model used to prioritize corporate-actions work at SIX Group. The model separates complexity (how hard a SWIFT corporate-action message is to process) from criticality (how urgent or impactful it is), then combines both into a single priority band (Very Low → Very High).
Context: Built for the corporate-actions operations team at SIX Group ZH. Coefficients can reflect an expert-tuned model or weights learned from historical data. Feature values are normalized to ([0, 1]); sub-scores are mapped to a 1–9 scale before combination.
Complexity — sliders, feature weights, and live LaTeX breakdown:
Criticality — position exposure, ISIN volume, due date, normalized score:
- Complexity tab — event type, security type, message length, market/sender complexity, foreign-market flag
- Criticality tab — position exposure, security criticality, ISIN market volume, due-date proximity
- Combined tab — weighted blend of complexity and criticality with LaTeX formula breakdown
- Transparent math — live LaTeX showing regression, min–max normalization, and rounding
For each dimension, the raw score is a linear combination (\beta_0 + \sum \beta_i x_i) with (x_i \in [0,1]). It is normalized against the min/max achievable with those coefficients, scaled to 1–9, and rounded. The combined score is round(w_complexity * s_complexity + w_criticality * s_criticality) and mapped to priority labels in the UI.