Wira Dhana Putra | Data & AI Portfolio
I turn messy data and unclear business problems into structured analysis, reproducible workflows, and deployed portfolio projects—with clear validation, transparent assumptions, and honest limitations
Selected work across analytics, data engineering, machine learning, and AI
Flagship AI/Data Quality Project
A Gemini-powered data quality copilot that combines a deterministic Python audit engine with read-only AI tool calling to help users inspect CSV datasets, calculate data quality scores, identify issues, and understand remediation priorities through a Streamlit interface.
Analytics & Data Engineering Project
An automated laptop market data pipeline that scrapes marketplace listings, cleans messy product specifications, tracks price changes with SCD Type 2 logic in SQLite, syncs processed data to Supabase/PostgreSQL on a schedule, and serves insights through a Streamlit dashboard.
Supporting AI Application
A multilingual, Gemini-powered message risk analyzer for suspicious text. It flags potential scams, hoaxes, and online gambling promotions while presenting AI-generated explanations as advisory signals that do not replace independent source verification.
Business Analytics & Statistics Case Study
A business analytics case study using funnel analysis, segmentation, regression, and comparative hypothesis testing to evaluate campaign performance and identify cost-efficiency patterns across observed audience groups. Findings are associative and do not claim randomized or causal impact.
Capabilities demonstrated through portfolio projects and prior business experience.
Exploratory analysis, funnel analysis, segmentation, regression, hypothesis testing, and business-focused interpretation.
Web scraping, ETL workflows, SQLite, Supabase/PostgreSQL, scheduled automation, and SCD Type 2 historical tracking.
Feature preparation, supervised modeling, evaluation, error analysis, and reproducible experiment workflows.
Gemini API integration, read-only tool-calling architecture, AI-assisted classification, and Streamlit-based AI interfaces.
Deterministic data checks, quality scoring, issue prioritization, readiness assessment, and limitation-aware reporting.
Streamlit dashboards, Power BI, Tableau, Excel, and clear communication of analytical findings.
Python application structure, Git/GitHub workflows, automated validation, scheduled jobs, and deployed demos.
Before moving deeper into Data and AI, I worked closely with operations, inventory control, reporting, and
process improvement. That experience taught me to look at problems not only from the data side, but also from
the business process behind the data.
Today, I bring that practical mindset into data projects: cleaning messy data, building reliable workflows,
analyzing patterns, validating results, and communicating insights in a way that can support better decisions.
Article about an automated laptop price pipeline with scheduled refresh, historical tracking, and dashboard deployment.
Read on MediumArticle about building an AI-assisted multilingual message risk analyzer and presenting its outputs with appropriate limitations.
Read on MediumCase study on funnel performance, audience segmentation, statistical comparison, and cost-efficiency patterns in observed groups.
Read on Medium
I combine business experience with technical execution across Data and AI. I turn unclear problems into
structured analysis and messy data into workflows that are easier to understand, test, and improve.
My portfolio focuses on practical, end-to-end projects across analytics, machine learning, data quality,
and AI applications. The work emphasizes transparent assumptions, reproducible methods, clear validation,
and honest limitations.
