[ DATA ANALYST × AI PRACTITIONER ]

Azizu Ahmad Rozaki

Data Analyst with 1+ years of experience building BI dashboards, automating reporting workflows, and driving data-informed decisions across multi-entity operations serving 20,000+ users. Published IEEE researcher (99% model accuracy) with hands-on expertise in Python, SQL, Power BI, and Tableau.

Semarang, Indonesia — Open to Relocate / Remote

Test Interactive SandboxView Case Studies
Published Researcher
IEEE
Operational Assets Tracked
Rp1.2T
Reporting Time Cut (8h → 3h)
-62%
Rookie of the Year 2025 (of 20)
#1
[ INTERACTIVE DEMO ]

Collection Risk & Recovery Simulator

A recovery-probability threshold trade-off, in the same shape as the scoring model behind my Collection Scoring Dashboard case study. Adjust the sliders to see precision, recall, and latency react live.

Illustrative simulation with synthetic data — not the real BPR production model or customer data.

collection_scoring_sim.py — illustrative
50%
50%
30%
PRECISION

69.5%

RECALL

70.5%

LATENCY

46ms

[ FEATURED WORK ]

Case Studies

Dashboards, models, and pipelines shipped at work or published as research — ordered most recent first.

01 / 03
Collection Priority Dashboard — sensitive figures blurred for confidentiality
[CREDIT-RISK / BI]

Collection Scoring Dashboard

BPR Gunung Rizki needed a systematic way to prioritize which delinquent borrowers (collectibility 3–5) to collect from first.

XGBoostFastAPITailwind CSSChart.js

BUSINESS IMPACT

  • → AUC-ROC 0.962 predicting 3-month recovery probability
  • → Powers daily collection prioritization across branches
Private — real customer data (BPR Gunung Rizki)
Tableau dashboard overview — revenue trend, regional and category breakdown
[RETAIL-ANALYTICS]

Store Performance & Churn Diagnosis

Aggregate sales figures across two stores masked a growing member-retention problem hidden in 159k+ transactions.

TableauSQL

BUSINESS IMPACT

  • → Uncovered 48% silent member churn (3,644 members)
  • → Flagged Rp466M in at-risk revenue
GitHub
Blinkit Power BI dashboard — sales, ratings, and outlet breakdown
[RETAIL-ANALYTICS]

Blinkit Store Performance (Power BI)

Aggregate sales data suggested Supermarket Type1 was outperforming — masking which store format was actually more efficient per outlet.

Power BIDAX

BUSINESS IMPACT

  • → Diagnosed outlet-count bias behind Supermarket Type1's sales lead
  • → Identified Grocery Store format as the true underperformer (-40% sales/outlet)
GitHub
[SQL-ANALYTICS]

Atliq Hardwares Sales Insights

Simulated management stakeholders needed regional and channel-level revenue insight buried in raw transactional data.

SQLCTEsWindow Functions

BUSINESS IMPACT

  • → Delivered market coverage & discount-trend insights
  • → Channel-level revenue contribution analysis
GitHub
[ CAPABILITIES ]

Technical Stack

Analytics & BI

SQL
Power BI
Tableau
Excel

Machine Learning & AI

Python
TensorFlow
PyTorch
Deep Learning (CNN)

Engineering & Tools

Git / GitHub
C++
Java
Figma
[ CREDENTIALS ]

Certificates

ADDITIONAL COURSEWORK — 8 certificates

[ TRACK RECORD ]

Experience

May 2025 — Present

Data Analyst

BPR Gunung Rizki — Semarang, Indonesia

  • Built and deployed 3 BI dashboards across 20 branches, giving management real-time visibility into Rp1.2T in operational assets.
  • Redesigned weekly reporting workflow, cutting preparation time from 8 to 3 hours while adding risk-exposure metrics.
  • Awarded "Best Rookie of the Year 2025" out of 20 employees.

Sep 2023 — Feb 2025

Teaching & Mentor Assistant

Bengkel Koding — Semarang, Indonesia

  • Mentored 50 students in Data Science, achieving 100% completion rate and an average assessment score of 89/100.
  • Authored 70%+ of curriculum content for Data Science and Deep Learning modules used by 3 cohorts.

Oct 2023 — Feb 2025

Research Assistant

DREAMS (Dinus Research Group for AI in Medical Science) — Semarang, Indonesia

  • Developed a CNN-based deep learning model (TensorFlow/PyTorch) for lymphoma sub-type classification, achieving 99% accuracy.
  • Published findings at the IEEE iSemantic Conference.

Apr 2024 — May 2024

Product Designer & iOS Developer

Apple Developer Academy — Surabaya, Indonesia

  • Designed and prototyped a wall-paint selection app in Figma, coded in SwiftUI/Xcode, as part of a 5-person team.
[ RESEARCH ]

Publications

Peer-reviewed work in applied machine learning, published while at DREAMS and Universitas Dian Nuswantoro.

[FIRST AUTHOR]

Lymphoma Sub-Type Classification Using DenseNet 169-Based Transfer Learning Architecture

IEEE iSemantic Conference — 2024

Read Paper
[CO-AUTHOR]

Accelerating Classification for IoT Attack Detection Using Decision Tree Model with Gini Impurity Tree-Based Feature Selection Technique

Jurnal Teknik Informatika (JUTIF) — Sinta 2 — 2025

Read Paper
[CO-AUTHOR]

Comparative Analysis of Vectorization Methods for Academic Supervisor Recommendations

Jurnal Teknologi Informasi dan Terapan (J-TIT) — 2024

Read Paper
[ CONTACT ]

Let's Build Something

contact.sh

$ whoami

Azizu Ahmad Rozaki — Data Analyst & AI Practitioner

$ cat location.txt

Semarang, Indonesia — Open to Relocate / Remote

$ cat email.txt

azizu.rozaki@gmail.com

$ ls ./profiles

github.com/azizuahmadlinkedin.com/in/azizuahmad