UJDC AI Integrated Financial Cockpit
Designed by David Kwan · CFO of UJDC
Cash Conversion Cycle
≈384d
vs benchmark 90-120d
Asset Turnover
<1.0×
Bottleneck identified
Aged Finished Goods
~HK$63m
≥3 years old
Live NPV
+0.66m
Positive
Pain-Point Dashboard
痛点仪表板Market Risk Monitor
市场风险监控Gold vs Diamond — Rebased to 100
Click any point to rebase.
Input-cost scissors: gold up +154%, diamonds down −30%
Customer Currencies vs USD — Rebased to 100
Click any point to rebase.
Revenue weights: AUD 35% · GBP 25% · EUR 20% · BRL 12%
CFO Decision Engine
CFO决策引擎Control Panel — Adjust Levers, See Impact
—
Cost of Equity
—
WACC
—
Risk-adj NPV
—
Payback
—
RA Benefit/yr
Scenario & Sensitivity
情景分析Explainable AI Lab
可解释AI实验室State-of-the-art ML models for gross-margin prediction. Each model's SHAP values decompose predictions into additive feature contributions — the CFO sees the "why", not a black box.
SHAP Results
Methodology & Equations
方法论Core Code Excerpts
## SHAP — Gradient Boosting + TreeExplainer
from sklearn.ensemble import GradientBoostingRegressor
import shap
model = GradientBoostingRegressor(n_estimators=400, max_depth=3,
learning_rate=0.04).fit(X_train, margin_dev)
explainer = shap.TreeExplainer(model)
shap_values = explainer.shap_values(X_test)## Causal AI — Back-Door Adjustment (Pearl, 2009)
from sklearn.linear_model import LinearRegression
import numpy as np
naive = LinearRegression().fit(M.reshape(-1,1), margin).coef_[0] # confounded
causal = LinearRegression().fit(np.c_[M, age], margin).coef_[0] # do(M)## Differential Evolution Optimiser
from scipy.optimize import differential_evolution
res = differential_evolution(neg_npv, bounds, maxiter=200, tol=1e-6, seed=42)