DATA SCIENCE · FRAUD ANALYTICS · CFE

Hi, I’m Shaneeza.

I’m a data scientist and Certified Fraud Examiner based in Washington, DC. I use data to understand financial crime and build tools that make complex information useful.

M.S. Business Analytics & AI, American University
B.S. Fraud Examination & Financial Forensics, John Jay College

My projects ↗My experience ↗Résumé ↓

PROJECTS

Selected projects

A few things I’ve built and explored. Open a project for the methods, results, and code.

FRAUD DETECTION · SNOWFLAKE + SQL

Turning 6.3 million transactions into a review queue

Behavioral risk scoring in Snowflake, from raw transactions to a prioritized fraud review queue.

The problem

PaySim contains millions of mobile-money transactions, but only 0.13% are fraud. A useful system had to find the risky behavior without treating every unusual payment as equally suspicious.

What I tried

I built raw, staging, and mart layers in Snowflake. Then I tested three behavioral signals: an origin account dropping to zero, a destination balance staying unchanged after receiving funds, and transaction types associated with every fraud case in the dataset.

What happened

The scoring rules sorted every transaction into four risk tiers. The HIGH tier contained 5,145 transactions, and 76% were labeled fraud. That gave an investigator a much smaller place to start.

What I’d test next

PaySim is synthetic. With real bank data, I would test stability over time, monitor feature drift, and measure how many legitimate customers the rules inconvenience.

Technical details on GitHub ↗View Demo ↗

Snowflake fraud detection pipeline — 6.3M rows · 3 behavioral features · 4 risk tiers

MARKET EVENT STUDY · PYTHON

What happens after corporate fraud becomes public?

An event study of 25 corporate fraud disclosures using the Fama-French three-factor model.

The question

I wanted to separate the market reaction to a fraud disclosure from normal market movement.

What I did

I curated 25 confirmed disclosures and used the Fama-French three-factor model to estimate expected returns before measuring abnormal returns around each event.

What I found

Cumulative average abnormal returns reached roughly −20% by day +10. The continued decline suggested that investors were still reassessing the company after disclosure day.

The limit

The sample is small and intentionally focused on high-profile cases. Other company news may overlap with some disclosure windows.

Methods and tests on GitHub ↗View Demo ↗

CAAR AFTER DISCLOSURE

Illustration of the event-study trend; see GitHub for the measured results.

DATA ENGINEERING · PYTHON + SQLITE

Building the table a fraud model needs

A modular ETL pipeline that cleans transaction data, builds fraud signals, and exports alerts.

The problem

Fraud demonstrations often begin with a clean file. Production work begins with duplicates, missing values, invalid amounts, and inconsistent records.

What I built

I generated 10,000 synthetic transactions, injected realistic data-quality problems, cleaned them, and created five fraud signals plus a composite risk score.

The result

The pipeline turned 10,100 raw rows into 9,753 validated records, a SQLite database, and a separate alert file. Twenty-seven tests check the workflow.

Why it is modular

The extract, transform, and load stages are separate, so a real API or warehouse can replace the synthetic source without rewriting the feature logic.

Pipeline and tests on GitHub ↗View Demo ↗

Fraud transaction ETL pipeline — 10,100 raw rows → 9,753 validated rows

DOCUMENT AI QA · PYTHON

Turning invoice extraction into a measurable test

A FastAPI evaluation workflow that validates invoice fields against ground truth.

The problem

A document extraction result can look right and still contain errors that would break a real workflow. I wanted a repeatable way to test every field instead of checking a few examples by hand.

What I built

I built a FastAPI service and Python client that submit invoices, poll jobs, validate the returned JSON, and compare each field with ground truth.

What I found

The baseline reached 45.8% field accuracy. After adding layout aliases and fixing a tax-label collision, the tuned configuration reached 100% on the three included synthetic layouts.

How I tested it

Twenty-one unit, contract, integration, and regression tests run through GitHub Actions. The dataset is small and synthetic, so the result tests the workflow, not production accuracy.

Tests and evaluation on GitHub ↗View Demo ↗

TRANSIT ANALYSIS · R

When two Metro stations need different plans

A comparison of weekday and holiday ridership at two Washington, DC Metro stations.

The problem

A transit schedule built around an average day can miss how demand changes by station, weekday, and holiday.

What I compared

I analyzed Monday and Friday ridership at Foggy Bottom-GWU and Tenleytown-AU from August 2021 through July 2024.

What stood out

Fridays were consistently busier, holiday Mondays dropped sharply, and Foggy Bottom had both higher volume and more variation than Tenleytown.

Analysis and visualizations ↗View Demo ↗

COMPENSATION ANALYSIS · R

Testing whether career level tracks salary in NYC

Standardized salary data from more than 5,300 NYC job listings to compare career levels.

The problem

Job titles are inconsistent, and posted pay can be hourly, daily, or annual. Comparing compensation required a common scale.

What I changed

I standardized wages, calculated midpoint salaries, grouped more than 5,300 listings into career levels, and tested the relationship with salary range.

The result

The association was statistically significant with a moderate effect size (Cramer’s V = 0.28). Leadership roles concentrated in the highest salary group.

Statistical workflow on GitHub ↗View Demo ↗

LABOR-MARKET ANALYTICS · R

Skills appear in bundles, not isolation

Association rules and statistical models exploring how skills, industry, and pay connect.

The question

Which skills appear together in LinkedIn postings, and how do those combinations relate to industry, experience, and salary?

What I tried

I joined job, skill, salary, and industry tables, then combined association rules, clustering, decision trees, and regression.

What emerged

Distinct skill-based job profiles had meaningful salary differences. Experience increased salary, but the size of that increase depended on industry.

Full analytical workflow ↗View Demo ↗

AUDIO ANALYTICS · R + QUARTO

What makes a song sound energetic?

A regression analysis of how loudness, tempo, and genre relate to Spotify track energy.

The question

I tested how loudness, tempo, and genre relate to the energy score assigned to Spotify tracks from the 2000s.

The approach

I cleaned the audio features, grouped genres, built a multiple regression model, and checked its assumptions and statistical power.

The finding

Loudness was the strongest predictor, while tempo had a smaller positive relationship. The model explained about 54% of the variation in energy.

Report and model diagnostics ↗View Demo ↗

BACKEND DEVELOPMENT · TYPESCRIPT + NODE.JS

Building a transaction risk assessment API

A TypeScript and Express REST API that validates transaction data and returns an explainable, rule-based risk assessment.

The problem

Transaction screening needs a consistent way to turn raw inputs into risk indicators that can be understood and reviewed.

What I built

I implemented a Node.js API with a health-check endpoint, runtime input validation, typed transaction models, and a scoring function that evaluates transaction amount, international status, and account age.

How it works

The API returns a score, LOW/MEDIUM/HIGH risk category, flag, and human-readable reasons. Invalid inputs receive a 400 response instead of being scored.

What I’d improve next

The current thresholds are illustrative rather than calibrated to observed fraud. Next steps include automated tests, transaction history, and evaluation against labeled data.

API source and documentation on GitHub ↗View Demo ↗

MACHINE LEARNING · PYTHON + SCIKIT-LEARN

Predicting survival with logistic regression

A Titanic survival classifier with interpretable features and a live browser-trained prediction model.

The question

Can passenger class, sex, age, family size, fare, and embarkation port help predict survival?

What I built

I cleaned missing data, encoded categorical features, and trained a logistic regression classifier with an 80/20 train-test split.

The result

The original Python model achieved approximately 80.9% training accuracy and 78.2% test accuracy. The live demo retrains a logistic regression model in the browser on the project's historical CSV and lets visitors test their own passenger profiles.

What to explore

Change ticket class, sex, age, fare, family size, or embarkation port and see how the estimated probability and feature contributions respond.

Model and preprocessing on GitHub ↗View Demo ↗

FINANCIAL CRIME · LIVE APPLICATION · JAVASCRIPT

Building a real-time fraud monitoring platform

An interactive fraud operations console that streams synthetic payments, calculates risk, and supports hands-on investigation.

The challenge

Investigators need to prioritize suspicious transactions as they arrive and distinguish risk signals from confirmed fraud.

What I built

A browser-executed transaction stream, transparent risk scoring, adjustable alert thresholds, manual transaction testing, and a review queue.

What visitors can do

Start or pause the stream, inject transactions, import labeled CSV data, investigate flagged activity, record review decisions, and export results.

Next iteration

Replace the heuristic engine with a trained model served through an API, add persistent case management, and monitor drift and feedback.

Application source on GitHub ↗View Demo ↗

EXPERIENCE

Experience

JAN–MAY 2026

Research Assistant

Schmidt Special Competitive Studies Project

Combined Airtable, Google Sheets, and third-party APIs into a company-research dashboard that cut meeting preparation time by 50%.
Automated sponsor-news collection, saving three hours of manual research each week.
Tracked actions across more than 30 sponsor companies and prepared weekly leadership summaries.

AUG 2024–DEC 2025

Fraud Data Analyst

EduGuide Overseas

Built anomaly-detection models in Python and R and classified more than 200,000 transactions by fraud risk each month.
Analyzed more than 500 payment datasets monthly and reduced suspicious student applications through automated scoring.
Built Power BI reporting that cut review time by 15%.

SUMMER 2024

Financial Crime & Investigative Services Intern

Guidehouse

Analyzed more than 20,000 ACH and wire transfers with SQL and Python.
Engineered transaction features and identified more than 50 high-risk patterns for compliance investigation.

SUMMER 2023

Fraud Audit Intern

New York State Office of the Attorney General

Used SQL and statistical validation to identify more than $500,000 in recovery anomalies.

EDUCATION

DEC 2026

M.S. Business Analytics & AI

Data Science and Statistics concentration
American University

MAY 2025

B.S. Fraud Examination & Financial Forensics

Computer Science minor
John Jay College

CERTIFICATES

Certificates

Certified Fraud Examiner (CFE)

Association of Certified Fraud Examiners (ACFE)

Issued Aug 2025

Google Analytics Certification (2026)

Skillshop

Issued Sep 2026 · Expires Sep 2027

Credential ID 193309294

UiPath Academy Agentic Automation Developer Associate Training

UiPath

Issued Jul 2026 · Expires Jul 2028

Credential ID 187576468

SQL Associate

DataCamp

Issued Jun 2026 · Expires Jun 2028

Credential ID SQA0017861460924

Microsoft Certified: Azure AI Fundamentals

Microsoft

Issued Jan 2026

Forensic Accounting and Fraud Investigations Course

Taxscan Academy

Issued Mar 2025

Credential ID 47403-ACA-2024

Networking Devices and Initial Configuration

Cisco Networking Academy

Issued Nov 2024

Networking Basics

Cisco Networking Academy

Issued Nov 2024

Commonwealth Bank - Introduction to Data Science Job Simulation

Forage

Issued Sep 2024

Credential ID E9k9nLk6SSdynZBD3

PwC Switzerland - Power BI Job Simulation

Forage

Issued Jul 2024

Credential ID bCzBaPBdtX4yCoirv

Goldman Sachs - Excel Skills for Business Job Simulation

Forage

Issued Jul 2024

Credential ID X4y7B9Fg63roWzLXD

Accenture North America - Data Analytics and Visualization Job Simulation

Forage

Issued Jul 2024

Credential ID XNNgXEuCeRPuNsDu3

Audit Virtual Experience Program

KPMG

Issued Jul 2023

Credential ID FccdhnJSMLXaoZXCq

Excel Essential Training (Office 365/Microsoft 365)

Issued Nov 2022

White Collar Defence & Investigations Virtual Experience Program

Latham & Watkins

Issued Nov 2022

Credential ID Az2ECD5v478tyrFsk

QuickBooks Online ProAdvisor Certification

Quickbooks Online Certified ProAdvisor

Issued Jul 2022

Microsoft Office Specialist: Microsoft Excel Expert (Office 2019)

Microsoft

Issued May 2022

Credential ID I252-4035

CONTACT

Let’s connect.

hasnani.shaneeza@gmail.com

LinkedIn ↗GitHub ↗Résumé ↓