Data / decision systems
Data analytics turns messy operational, product, and financial data into cleaned models, dashboards, executive summaries, and business decisions.
What this system does
Cleans raw data, standardizes fields, handles missing values, removes duplicates, and creates reusable summary tables.
Builds SQL and Python workflows that support dashboards, reporting automation, QA checks, and executive readouts.
Connects analysis outputs to stakeholder questions, product prioritization, and operational planning.
How I use it
Portfolio analytics SQL demonstrates messy sample data, cleaning logic, KPI summaries, product analytics queries, and cleanup.
BudgetDB models software spend, headcount allocation, and executive reporting views in Postgres.
Python churn script shows a reproducible modeling workflow with target conversion, missing-value handling, encoding, train/test split, RMSE, and AUC.
Related projects
Evidence or examples
portfolio-analytics-demo.sql includes messy raw data, cleaning pipeline, KPI summary tables, product analytics queries, and safe cleanup.
software-cost-allocation.sql creates Postgres views for employee counts, company software cost per employee, and allocated team cost.
predictive-churn-analysis.py loads data, transforms target labels, handles missing values, encodes categorical fields, and reports RMSE/AUC.
CASE
WHEN quantity ~ '^\d+$' THEN quantity::integer
WHEN lower(quantity) = 'two' THEN 2
ELSE NULL
END AS quantitySELECT sale_date, COUNT(DISTINCT customer_id) AS unique_customers,
SUM(quantity) AS total_items_sold,
ROUND(SUM(net_revenue), 2) AS total_revenue
FROM clean_daily_sales
GROUP BY sale_date;X_train, X_test, y_train, y_test = train_test_split(
X, y, test_size=0.2, random_state=42, stratify=y
)