With a background in Product Management and Analysis, I bridge the gap between technical data systems and business operations. I focus on making effort to boost productivity, cut operational costs, mitigate compliance risks, and accurately forecast resource demands, ensuring the sustainable enterprise growth.
60-hour end-to-end analysis on 100k+ orders from the Olist marketplace.
Applied K-Means clustering to segment sellers into risk tiers, identified
freight margin leakage, and built a delivery-delay satisfaction intelligence
layer delivered as a Power BI executive dashboard.
Insight: Seller Risk Scoring Requires Both Dispatch Metrics and Customer Voice. NLP
enrichment from review text adds a layer of early warning that SLA-only scoring cannot capture. Freight
Leakage Lives in the Long Tail, Not the Average.
Predicting customer churn for a telecommunications provider using
historical subscription and usage data. Logistic Regression and Random
Forest models identify high-risk customer segments, enabling proactive
retention campaigns before churn events occur.
Insight: Predicting Churn Requires Behavioral Signals, Not Just Demographics. The
strongest predictors in telecommunications are behavioral aspects like contract tenure and service bundle
counts.
Cohort-level analysis of 32,593 student records from the Open University
Learning Analytics Dataset (OULAD). Identifying engagement drop-off patterns,
assessment submission behaviors, and predictors of student withdrawal to
support early intervention strategies.
Insight: Non-submission is a withdrawal announcement — 92% of students who miss
the first assessment deadline exit the course. VLE engagement diverges as early as week 3,
giving a 6-week window to intervene before withdrawal becomes irreversible.
Swinburne University of Technology, Australia
Benedictine University (USA), Vietnam Cohort
Available for entry-level to graduate positions in data analysis, product analysis, business analysis, and associate machine learning roles in Australia.