Kevin Ngo

Data Analyst Intern @ Capital One

Jun 2025 - Aug 2025

Reduced retrieval time of historical credit bureau reports from minutes to seconds by onboarding a monitoring SQL script to an internal platform, improving report accessibility for stakeholders managing 60K+ daily customer cases
Established an auditable, documented script review process, by registering the script to a governance platform, eliminating fragmented communication between the data analyst and credit bureau disputes teams
Prevented data-syncing issues that would have resulted in inaccurate results for 3 daily reports used by the credit bureau disputes team by identifying upstream data changes and correcting the reporting scripts on Databricks

March 2025 - Aug 2025

Built a comprehensive, nationwide dataset of 2M+ alcohol retailers as part of a CDC-funded research initiative, supporting an analysis evaluating the effectiveness of firearm-restricting policies across 3K+ counties
Reduced data cleaning time from 1 day to 3 hours per state by designing data preprocessing scripts using R, and created clear onboarding documentation for new student researchers
Mapped county-level disparities in income, education, and demographics using R to support research on whether policies restricting firearm access near alcohol-serving venues reduce gun violence

June 2024 - August 2024

Analyzed conversion rates from a 100K-customer marketing experiment using SQL and convinced Marketing to increase the sample size of future experiments, improving the chance of detecting meaningful conversion rate lifts to 80%
Quantified uncertainty in conversion rate estimates by creating 70% bootstrap confidence intervals using Python, helping stakeholders recognize the inconclusive findings and avoid incorrect interpretations of the results of the experiment
Built a SQL query that joined Safeco Prospecting Program data from a vendor with Liberty Mutual's internal data using fuzzy-matching techniques, identifying ∼ 60% more prospects compared to the previous method