Willsen Wijaya — Data Analyst & Engineer

About me

I'm a recent Information Systems graduate from Universitas Multimedia Nusantara (GPA 3.91/4.0), specialising in Big Data Analytics. My focus sits at the intersection of data analytics, engineering, and machine learning — building systems that don't just report the past, but help organisations act on the future.

Over the past year I've interned across two divisions at PT Bank Central Asia (BCA) — one focused on customer analytics and automation, the other on ETL infrastructure and internal reporting platforms. I've shipped production-grade dashboards, ETL pipelines serving 5 business domains, and API integrations used by bank employees daily.

Outside of work, I research predictive modelling for financial assets. My thesis explored Bitcoin and Ethereum price prediction using LSTM, GRU, and XGBoost with macro-financial indicators — the best model (XGBoost) achieved a MAPE of 2.18% for Bitcoin and was deployed as a live Streamlit application.

Here are some technologies I work with regularly:

Python SQL Power BI Tableau SSIS / SSRS Power Automate SAS Visual Analytics ASP.NET OutSystems

Education

B.S. Information Systems
Universitas Multimedia Nusantara
2022 – 2026 · GPA 3.91 / 4.0

Specialisation: Big Data Analytics
Relevant: Database Systems, ETL, Business Intelligence, Data Visualization, Data Analytics, Machine Learning, Deep Learning

Certifications

IBM Python for Data Science, AI & Development Nov 2025

PASAS® CISDM – Certified International Specialist in Data Modelling Mar 2026

Core focus areas

Data Analysis & Visualization
ETL & Reporting Automation
Machine Learning & Deep Learning
Dashboard Development
Time-Series Forecasting

Where I've worked

Data Analyst Intern · PT Bank Central Asia (GSIT – MIS)

Aug 2025 – Feb 2026 · Jakarta, Indonesia

Built end-to-end automated ETL pipelines (SSIS + SQL) covering extraction, transformation, staging, validation, and automated report distribution — deployed across 5 business reporting domains: collateral, loans, KUR, intraday liquidity, and beehive.
Transformed manual report generation into self-service web-based reporting platforms via SSRS, allowing stakeholders to access up-to-date reports through web links with automated data refreshes — no manual intervention needed.
Built ASP.NET API endpoints integrated into an OutSystems internal application, supporting FAQ workflows, BPRO request tracking, monitoring, and role-based information access for bank employees.
Designed and deployed SSRS reports with standardised tabular layouts, parameters, and formatting for multiple business units.

Data Analyst Intern · PT Bank Central Asia (CDG – OSS)

Feb 2025 – Jul 2025 · Jakarta, Indonesia

Engineered a dynamic Customer Survey Dashboard (CSI, NPS, CSAT) using Power BI with Python for data preprocessing — replacing manual Excel-to-PowerPoint workflows with drill-through, period-comparison analytics for the executive management team.
Automated recurring data extraction from HaloReport using Power Automate + Python, reducing manual retrieval time from 3–4 hours to under 1 hour.
Delivered data-driven insights on customer pain points that directly accelerated decision-making velocity for senior leadership.

Things I've built

★ Featured project / Thesis Cryptocurrency Price Prediction — BTC & ETH

Multivariate time-series forecasting for next-day Bitcoin and Ethereum prices using LSTM, GRU, and XGBoost, with macro-financial indicators (Gold, US Dollar Index, S&P 500). Two feature scenarios were evaluated; the best-performing models were deployed in an interactive Streamlit application with a forecast log and model info panel.

Included Wilcoxon signed-rank statistical significance tests and SHAP feature importance analysis to validate model quality and interpretability.

Python LSTM GRU XGBoost TensorFlow Keras SHAP Streamlit

View live app ↗

Model performance

BTC best model XGBoost + Macro

BTC MAE 1,804.67

BTC MAPE 2.18%

ETH best model GRU (historical)

ETH MAE 84.26

ETH MAPE 3.19%

Models compared 6 (3 algo × 2 feat)

Deployment Streamlit (real-time)

★ Academic Project / Publication · 2024 Waste Image Classification — MobileNetV2

CNN model classifying 12 categories of household waste using MobileNetV2 with fine-tuning on a 6,000+ image dataset. Achieved 85.56% accuracy — outperforming vanilla MobileNetV2 (80.49%). Deployed as a real-time Streamlit web application.

TensorFlow MobileNetV2 Transfer Learning Streamlit

View live app ↗

Model performance

Best model MobileNetV2 (fine-tuned)

Base accuracy 80.49%

Tuned accuracy 85.56%

Classes 12 household categories

Dataset 6,000+ images

Deployment Streamlit

★ Interactive Web Application Office Equipment Mapping System

A web-based application built to visually map and manage office equipment on a digital floor plan. Features include drag-and-drop device placement on interactive canvas elements, real-time database synchronization, and Excel data export capabilities.

React.js Konva.js Firebase Firestore SheetJS (Excel)

⚠ Data shown is for demonstration purposes only.

BCA – GSIT Division

ETL & Reporting Pipeline — SSIS / SSRS

End-to-end ETL pipelines using SSIS covering extraction, transformation, staging, and final table creation across 5 business domains. Automated report distribution via email and self-service SSRS web portal with standardised layouts and parameters.

SSIS SSRS SQL Server ASP.NET OutSystems

BCA – GSIT Division

MIRA Job Tracking Application & API

Developed REST API endpoints using C# on ASP.NET and integrated them into MIRA, an internal OutSystems-based application. The system supports job request (BPRO) tracking, progress monitoring for the MIS division, and role-based FAQ management.

ASP.NET C# OutSystems REST API SQL Server

BCA – CDG Division

Customer Survey Dashboard — Power BI

Interactive Power BI dashboard visualising telesurvey data (CSI, NPS, CSAT) for BCA's executive management. Python preprocessing pipeline with drill-through views and period-comparison analytics — replaced an entirely manual Excel-to-PowerPoint workflow.

Power BI Python DAX Data Preprocessing

BCA – CDG Division

Data Retrieval Automation — Power Automate

Built a Robotic Process Automation (RPA) pipeline using Microsoft Power Automate and Python to automate repetitive report downloads from the internal HaloReport system, reducing data retrieval time from 3–4 hours to under 1 hour (over 70% faster) while eliminating manual extraction and improving reporting efficiency.

Power Automate Python RPA Data Extraction

Academic Project · 2024

Banking Customer Churn Analysis

Developed and compared six machine learning models (Gradient Boosting, Random Forest, Decision Tree, SVM, Neural Network, Bayesian Network) to predict customer churn. Gradient Boosting achieved the best accuracy at 86.88% and identified Age as the strongest churn predictor.

Python Gradient Boosting Scikit-learn Feature Analysis

Academic Project · 2024

Big Data Analytics — Bank Deposit Subscription

Applied predictive, prescriptive, text, and sentiment analytics using the DCOVA framework to identify key drivers of bank deposit subscriptions. Built forecasting, machine learning, and NLP pipelines, achieving 86.02% classification accuracy with Gradient Boosting.

SAS Python NLP DCOVA Framework

Willsen Wijaya.

I turn data into decisions.

About me

Where I've worked

Things I've built

Let's build something together.