Primary Deliverables
Technical Ebook
12-chapter deep-dive covering the full data platform: architecture, Medallion pipeline, ML models, fintech integration, SARB regulatory reporting, and AML compliance.
12 ChaptersChart.jsSelf-contained
Interactive ERD
6-tab Mermaid entity-relationship diagram covering all Gold schema domains: Core Banking, Lending, AML, Cards & Fintech, Treasury, and Regulatory. Includes zoom/pan controls.
6 DomainsMermaid.jsZoom/Pan
Merchant Intelligence Map
Interactive Leaflet.js map of SA merchant transaction volumes and fraud hotspots. Toggle between Sales Volume, Fraud Heatmap, and AML Case views across 9 provinces.
Leaflet.jsSA ProvincesLive Filters
Executive Presentation
15-slide investor-grade deck: platform overview, KPIs, ML model performance, financial highlights, regulatory posture, and fintech ecosystem integration.
15 SlidesChart.jsExec-ready
Data Dictionary
Searchable column-level documentation for all 15 Gold schema tables: data types, primary/foreign keys, nullable constraints, and business descriptions with SA regulatory context.
15 TablesSearchableAll Layers
Analytics Workbook
Excel financial model with KPI dashboards, loan portfolio analysis, fraud trend charts, fintech revenue breakdown, and branch performance summary. Dynamic formulas throughout.
ExcelDynamicKPI Dashboards
Medallion Pipeline Architecture
🥉
BRONZE
Raw Ingestion
ADLS Gen2
ADLS Gen2
›
🪙
SILVER
Cleansed & Typed
Delta Tables
Delta Tables
›
🥇
GOLD
Star Schema
15 Dim + Fact Tables
15 Dim + Fact Tables
›
🤖
ML LAYER
12 Models
MLflow Tracking
MLflow Tracking
›
📊
CONSUMPTION
Power BI + dbt
Regulatory Reports
Regulatory Reports
Machine Learning Models
Credit Scoring
0.847
AUC-ROC
XGBoost · 500K customers
Fraud Detection
0.921
AUC-ROC
Random Forest · Real-time
AML Risk Scoring
0.889
AUC-ROC
Gradient Boost · FIC Act
Churn Prediction
78.3%
Accuracy
Logistic Regression
Loan Default (PD)
0.832
AUC-ROC
XGBoost · IFRS 9 ECL
CLV Segmentation
0.74
Silhouette Score
K-Means · 5 Clusters
NLP Sentiment
83.1%
F1 Score
BERT · Complaints
Tx Anomaly Detection
0.876
Precision@K
Isolation Forest
Code & SQL Assets
Databricks Notebooks (8)
Bronze ingestion → Silver transformation → Gold star schema → ML credit scoring → Fraud detection → AML risk → Regulatory reporting → Pipeline orchestration.
PySparkDelta Live TablesMLflow
SQL DDL Scripts
Bronze, Silver, and Gold schema DDL. Full Gold star schema with DIM/FACT tables, fintech settlement model, and seed data for all reference dimensions.
DDLDMLViews
dbt Models (14)
Staging, intermediate, and mart layers. Includes customer 360, credit risk, fraud intelligence, regulatory capital adequacy, and treasury analytics marts.
dbt CoreJinjaTests
Data Generator
Python script generating 500K customers, 12M+ transactions, 280K loans, fraud alerts, card transactions, SA fintech settlements, and AML cases — all SA-realistic.
PythonPandasFaker
SA Fintech Ecosystem Partners
Yoco · FT001
POS Card Acquiring. MDR 2.95%. ~250K active merchants. R12.4B annual volume. T+1 settlement. Raised $83M Series C.
Low RiskSME Focus
SnapScan · FT002
QR Code Payments. MDR 2.75% (lowest). ~180K merchants. R8.7B volume. T+1 settlement. Standard Bank subsidiary since 2014.
Low RiskQR Payments
PayFast · FT003
E-commerce Gateway. MDR 3.50% + R2.00 min. ~95K merchants. R15.2B volume. T+2 settlement. DPO Group member.
Medium RiskE-commerce
Peach Payments · FT004
API Payment Platform. MDR 2.90% + R1.50 min. ~32K merchants. R6.8B volume. T+1 settlement. API-first mid-market.
Low RiskAPI-first
PayGate · FT005
Enterprise Payment Gateway. MDR 2.80% + R1.00 min. ~18K enterprise merchants. R22.1B volume. T+2 settlement. DPO Group.
Medium RiskEnterprise