Aditya Khetarpal

Went into 816 historical credit profiles with no assumptions and used advanced SQL aggregations to reverse-engineer how underwriting decisions were actually being made. Found a hard credit score floor of 651 and a DTI ceiling of 39% that no one had explicitly documented but were clearly embedded in every approval.

Turned that finding into a vectorised Python rule engine that evaluated the same compliance logic in under a second, replacing a review process that had previously taken multiple business days. Every output came with a standardised, audit-ready log.

Built a Tableau dashboard that made the findings accessible to non-technical decision-makers, showing exactly where applications sat relative to the hidden thresholds and what the distribution looked like across the full dataset.

Processed over one million transaction rows using SQL and Pandas to build clean, validated customer profiles scored across Recency, Frequency, and Monetary dimensions using algorithmic quantile-scoring. The pipeline was designed to be reproducible and scalable, not a one-off analysis.

Discovered that the revenue distribution followed a sharp Pareto pattern across $17.74M in total revenue, and identified a $2.3M at-risk band tied to 2,400 customers whose engagement was actively declining. That became the basis for a targeted retention strategy.

Built an interactive Power BI dashboard with custom DAX measures that gave the business team a self-serve view of customer lifecycle tiers, so they could answer their own questions without coming back to me every time the data changed.

Analysed over 1,400 employee records to figure out not just who was leaving, but why. Ran exploratory analysis across compensation, working hours, department, and tenure to isolate which variables were actually driving attrition versus which ones just looked correlated.

Found that employees working overtime were leaving at 3x the rate of those who weren't, and that departing employees were earning on average $2,045 less per month than retained staff in comparable roles. Two numbers that told the whole story.

Delivered a Tableau dashboard that let HR leadership slice the data by team, seniority, and time period, alongside a structured recommendation deck that translated the analysis into a specific compensation strategy with an estimated cost of intervention.

Built a live data pipeline that pulls real-time PM2.5 air quality readings across five cities from a public API, with automated retry logic and failure handling built in from the start. If the API goes down, the pipeline recovers without manual intervention.

Cross-validated incoming data against a deliberately corrupted 44-row synthetic dataset and wrote automated logic that flags 25% of records as missing or anomalous before they reach analysis. Data quality is enforced at the pipeline level, not caught after the fact.

Applied SQL window functions (RANK/DENSE_RANK) and trend analysis to surface a 136.4 µg/m³ breach that was 9x the WHO safe threshold. Built a Power BI monitoring dashboard that displays risk levels in real time with documented metric definitions and data lineage, so anyone picking it up can understand exactly where every number comes from.

Built and continue to maintain a live financial analytics platform that tracks over $16 billion in Solana stablecoin supply across USDC, USDT, PYUSD, and USDG, with automated pipelines refreshing every five minutes from multiple external data sources.

Used Python's Requests library for API integration and wrote clean, reusable transformation scripts that handle data ingestion, validation, and loading into a structured analytics layer. Every stage of the pipeline is documented so it can be picked up and maintained by someone else.

Built dashboards tracking peg health, DeFi TVL, and protocol-level breakdowns that institutional researchers use for quarterly analysis.

Validated ability to configure and manage Google Analytics properties, interpret data across reporting workspaces, and translate business objectives into measurable events and conversions. Demonstrates practical understanding of attribution models, audience building, and cross-channel data activation for data-driven marketing decisions.

Covers the principles and techniques behind building clear, effective data visualisations using tools including Tableau and IBM Watson. Demonstrates the ability to communicate complex analytical findings through well-designed charts, dashboards, and presentations that drive informed decision-making.

Covers foundational statistical methods including measures of central tendency, variability, probability distributions, hypothesis testing, and statistical inference. Demonstrates the ability to apply both descriptive and inferential techniques to analyse datasets and draw meaningful, evidence-based conclusions.

Covers structured approaches to categorising and labelling data based on sensitivity, type, and business context. Demonstrates understanding of data classification frameworks, analytical skill application, and how proper data organisation supports governance, quality, and downstream analysis.

Introduces core concepts in data analysis and data visualisation, covering how data is collected, processed, and presented to generate business value. Demonstrates a foundational understanding of the data lifecycle and the tools and techniques used to turn raw data into actionable insights.