Frame the decision
Start with the operational question, the source of truth, and the point at which a person needs confidence to act.
Data systems, analytical clarity, and useful automation.
Explore Work ↓Data / Design / Decisions
01 / About
Data analyst focused on transforming raw signals into actionable decisions, engineering resilient automation pipelines, and designing high-clarity analytical dashboards.
02 / Selected systems
Explore focused work in automation, analytics, and machine learning. Each project traces a practical path from source data to a useful decision layer.
Predictive ML & Dispatch Economics
SignalDual-stage ML delay risk engine with zero-leakage pipeline and dynamic threshold economics models $8.5M net operational savings across 58k validation flights ($25.5M/yr annualized).
Financial Crime Surveillance
SignalDemonstrates how SQL window functions and multi-page Power BI dashboards bridge the gap between raw unstructured transaction logs and audit-ready anti-fraud operations.
Commercial Underwriting & Basel II Scorecard
SignalEnd-to-end 5C credit appraisal and Basel II scoring engine evaluating 150k debtor records with macro commodity sensitivity and IFRS 9 Expected Loss modeling.
Interactive 3D WebGL Studio
Signal3D WebGL airspace network visualizes ground taxi elevation pillars and great-circle late-aircraft ripple propagation across 7.08M commercial flights.
Interactive 3D Surface Studio
SignalInteractive 3D manifold topography visualizes 35.5 years of non-Gaussian geopolitical oil shocks.
Interactive 2D Operations Cockpit
SignalInteractive operations cockpit isolates late-aircraft ripple propagation and surface runway bottlenecks across 7.08M commercial flights.
Interactive 3D Graph Studio
SignalInteractive 3D Force-Directed Knowledge Graph exposes coordinated money mule funnels and credential stuffing clusters invisible in flat 2D tabular logs.
Global Health Surveillance
SignalDeconstructed 30 years of global cancer registries across 204 sovereign nations.
Interactive 3D Manifold Studio
SignalInteractive 3D Decision Hyperplane slices through latent transaction feature space, dynamically maximizing blocked fraud capital while minimizing customer friction.
Supply Chain & RFM Matrix
SignalDiagnosed cross-state freight bottlenecks and isolated high-ticket retention segments across 99.4k Brazilian orders.
Masterclass Data Architecture & Mesh
SignalPanduan belajar interaktif arsitektur Data Mesh: memecah data monolith menjadi 5 data product otonom dengan kontrak ODCS dan kepatuhan UU PDP.
Payment & Installment Engine
SignalProved credit installments drive 3.3x larger baskets and diagnosed the 10x checkout default anomaly.
Energy Risk & Econometrics
SignalDecoded 35.5 years of crude oil volatility, quantifying the impact of 7 global geopolitical crises.
Dataset + model snapshot
SignalStatic EDA, NLP, and browser-only inference.
5-Model Hybrid Ensemble
Signal5-model ensemble with active learning reduced multi-brand SKU reconciliation from days to minutes.
Financial Reconciliation Engine
SignalAutomated cross-document reconciliation slashed DBO month-end closing from 1 week to 1 day.
Batch Generation Engine
SignalTransformed a 4-hour manual copy-paste workflow into a 25-second automated batch compilation.
03 / How I work
Start with the operational question, the source of truth, and the point at which a person needs confidence to act.
Normalize inputs, surface uncertainty, and keep the path from raw data to output reviewable.
Turn the analysis into a repeatable tool, dashboard, or workflow that reduces avoidable manual work.
04 / Skills & tools
Automation
Analytics
Machine learning
Systems
05 / Real-world data playground
Real panel dataset (1990–2019) from Our World in Data & IHME Global Burden of Disease. Inspect 30-year epidemiological trajectories, longitudinal shifts, and mortality distribution across major cancer classifications.
7,401,940 recorded deaths in 2019 (all types).
Net growth from 1990 (4,315,874 baseline deaths).
1990–2019 standardized longitudinal panel (228 entities).
Mortality Trajectory (1990–2019)
Global Total (in thousands)Annual Records (30 Years)
| Year | Classification | Deaths (k) | Exact Count |
|---|---|---|---|
| 2019 | Global Total | 7,402k | 7,401,940 |
| 2018 | Global Total | 7,188.6k | 7,188,587 |
| 2017 | Global Total | 6,992.9k | 6,992,721 |
| 2016 | Global Total | 6,851.3k | 6,851,377 |
| 2015 | Global Total | 6,705k | 6,704,942 |
| 2014 | Global Total | 6,553.3k | 6,553,280 |
| 2013 | Global Total | 6,439k | 6,438,966 |
| 2012 | Global Total | 6,343.6k | 6,343,613 |
| 2011 | Global Total | 6,250.6k | 6,250,608 |
| 2010 | Global Total | 6,159.4k | 6,159,571 |
| 2009 | Global Total | 6,048.9k | 6,048,865 |
| 2008 | Global Total | 5,952.2k | 5,952,260 |
| 2007 | Global Total | 5,831.2k | 5,831,172 |
| 2006 | Global Total | 5,748.8k | 5,748,774 |
| 2005 | Global Total | 5,715.2k | 5,715,123 |
| 2004 | Global Total | 5,613.7k | 5,613,823 |
| 2003 | Global Total | 5,527.8k | 5,527,822 |
| 2002 | Global Total | 5,428.8k | 5,428,709 |
| 2001 | Global Total | 5,326k | 5,325,983 |
| 2000 | Global Total | 5,243.6k | 5,243,498 |
| 1999 | Global Total | 5,134.3k | 5,134,217 |
| 1998 | Global Total | 5,014.9k | 5,015,018 |
| 1997 | Global Total | 4,928.1k | 4,927,913 |
| 1996 | Global Total | 4,862.6k | 4,862,490 |
| 1995 | Global Total | 4,800.4k | 4,800,478 |
| 1994 | Global Total | 4,722.4k | 4,722,277 |
| 1993 | Global Total | 4,628.3k | 4,628,238 |
| 1992 | Global Total | 4,498k | 4,498,065 |
| 1991 | Global Total | 4,401.9k | 4,401,864 |
| 1990 | Global Total | 4,315.8k | 4,315,874 |
Data Provenance: Source records from the Global Burden of Disease Study (IHME / Our World in Data). Processed locally into static-first JSON with zero external API latency, demonstrating transparent analytical exploration and reproducible aggregation.
Open Full Working Portfolio Study ↗06 / Where I've been
Now
Building reliable analytical workflows and automated reconciliation systems.
Focus: financial automation, multi-model SKU mapping, and decision-ready dashboards at PT. Depoguna Bangunan Online.
2023 — 2024
Developing transaction anomaly detection engines and commodity risk models.
Focus: financial crime surveillance, time-series volatility, and SKKNI national competency standards.
Earlier
Exploring practical automation, lightweight tools, and reproducible systems.
Focus: reducing manual friction with clear inputs, simple logic, and fast execution.