👉 Open the Live SQL Terminal to execute custom queries and export results to CSV.

• Business Problem: Evaluate delivery lead-times across Brazilian states, detect logistical bottlenecks, and analyze customer repeat purchasing trends.
• Methodology: Queried 99,000+ customer records across 6 relational tables using Google BigQuery; wrote modular Common Table Expressions (CTEs), multi-table joins, and ranking window functions.
• Key Finding: Identified a 3.5x regional delivery lead-time disparity between northern and southeastern states, and demonstrated how expanding regional fulfillment centers would protect a 137% YoY revenue surge.
• Key SQL Pattern:

  WITH delivery_metrics AS (
    SELECT customer_state, AVG(DATE_DIFF(order_delivered_customer_date, order_purchase_timestamp, DAY)) as avg_days
    FROM `target.orders` JOIN `target.customers` USING(customer_id)
    GROUP BY customer_state
  )
  SELECT customer_state, avg_days, DENSE_RANK() OVER(ORDER BY avg_days DESC) as rank
  FROM delivery_metrics;

• Direct Project Links:
  🔗 GitHub Repository • 📄 Master SQL Script (all_target_queries.sql) • 📁 SQL Queries Directory (7 Scripts) • 📊 Case Study PDF Report

• Business Problem: Test demographic spending assumptions across 550,000+ Black Friday transactions to guide seasonal stock planning and marketing budget allocation.
• Methodology: Applied Central Limit Theorem (CLT) sampling distributions ($n \in {30, 100, 500, 1000}$), calculated standard error contraction, and built 90%, 95%, and 99% Confidence Intervals across demographics.
• Key Finding: Men spend +$702.96 (+8.05%) more per transaction overall ($9,437 vs. $8,734) with zero CI overlap at 99% confidence (gap >$649), yet within specific categories (such as Product_Category 1), spending between men and women is virtually identical ($11 difference). Marital status showed no spending variance ($4.74 difference, completely overlapping CIs).
• Key Python Snippet:

  # Computing sample mean standard error and 95% confidence interval
  std_err = df['Purchase'].std() / np.sqrt(len(df))
  margin_error = stats.norm.ppf(0.975) * std_err
  ci_95 = (df['Purchase'].mean() - margin_error, df['Purchase'].mean() + margin_error)

• Direct Project Links:
  🔗 GitHub Repository • 📓 Jupyter Notebook • 📊 Case Study PDF Report • 📱 LinkedIn Carousel Slides

• Business Problem: Define distinct customer profiles across entry-level (KP281), mid-tier (KP481), and commercial (KP781) treadmills to target sales recommendations.
• Methodology: Constructed two-way contingency tables, evaluated conditional probabilities $P(\text{Product} \mid \text{Fitness Rating})$ vs. demographic indicators, and conducted distribution analysis.
• Key Finding: Self-rated fitness (level 4–5) was the single strongest indicator for commercial model purchases, showing an 80%+ conditional probability for the premium KP781 model, proving that fitness habits matter far more than age or income alone.
• Key Python Snippet:

  # Conditional probability matrix
  pd.crosstab(index=df['Fitness'], columns=df['Product'], normalize='index') * 100

• Direct Project Links:
  🔗 GitHub Repository • 📓 Jupyter Notebook • 📊 Case Study PDF Report • 📈 Tableau Packaged Dashboard (.twbx)

• Business Problem: Leadership needed immediate visibility into loss-making product categories, regional discount sensitivity, and customer margin contributions.
• Methodology: Built a dynamic parameter-driven Tableau Public dashboard with interactive filters for Region, Segment, and Category; engineered Level of Detail (LOD) expressions and custom profit margin calculations.
• Key Finding: Tracked $733K+ in total sales and discovered that aggressive discounting on Tables in the Central region accounted for over 60% of total regional operating losses, providing the justification for discount limits.
• Direct Project Links:
  🌐 Live Interactive Dashboard (Tableau Public) • 📊 Executive Dashboard PDF Summary • 👤 Tableau Public Profile

• Operational Reality: Daily monitoring of sales volumes, physical tank stock levels, and dealer credit exposure across 350+ petroleum retail outlets.
• Methodology: Designed structured Excel reporting templates connecting SAP ECC ERP extracts, dynamic Pivot Tables, and automated VBA macros; assessed 2-year sales trends to calculate bank credit eligibility under the e-DFS facility for 50+ retail dealers.
• Quantified Impact: Reduced daily MIS compilation from 2.5 hours down to 25 minutes (70% reporting time saved) and coordinated documentation and operational compliance for the commissioning of 37 new retail outlets.

• Business Problem: Analyze driver allocation patterns, morning peak-hour ride cancellations, and customer retention drop-offs across urban pickup zones.
• Methodology: Solved 15 distinct business scenarios using window functions (ROW_NUMBER, RANK), multi-condition aggregations, and distance-based user segmentation.
• Direct Project Links:
  🔗 GitHub Repository • 📁 15 SQL Queries Directory

• Business Problem: Analyze department-level turnover risks, managerial spans of control, and salary equity across regional offices.
• Methodology: Wrote 9 structured queries utilizing correlated subqueries, window percentiles, and group-level benchmarking against company-wide averages.
• Direct Project Links:
  🔗 GitHub Repository • 📁 9 SQL Queries Directory

• Business Problem: Analyze machine telemetry data (operating temperatures, vibration spikes, cycle rates) across manufacturing lines to isolate unscheduled downtime.
• Methodology: Processed industrial IoT telemetry logs, synthesized failure correlations, and built an executive machine health dashboard.
• Direct Project Links:
  📜 Deloitte Australia Completion Certificate • 📈 Daikibo Telemetry Dashboard (.twbx) • 🖼️ Dashboard Screenshot Preview • 📊 Forensic Technology Equality Analysis (Excel)

• Business Problem: Examine catalog shifts between movies and television series over a 15-year period, international content growth, and director networks.
• Methodology: Handled nested genre strings, performed temporal trend analysis, and visualized release patterns using Pandas, Matplotlib, and Seaborn.
• Direct Project Links:
  🔗 GitHub Repository • 📓 Jupyter Notebook • 📊 Case Study PDF Report