Associate | Data & Analytics
PwC India · Jul 2021 – Present
- Engineered Azure Synapse pipelines to ingest CSV data from ADLS(Gen2), with pre-load validation logic verifying expected file counts and naming conventions before triggering transformation - eliminating manual checks and reducing pipeline failure incidents by ~60%.
- Built parameterized pipelines and mapping data flow transformations for cleansing, business-rule application, and structured loading into Azure SQL source tables; architected 100+ KPI-specific mapping data flows each performing table joins with complex filter and aggregation logic - reducing KPI computation time by ~40% vs prior manual processes.
- Orchestrated end-to-end pipeline execution using event trigger, ensuring analysis-ready KPI data reached the Power BI reporting team.
- Implemented a Bronze → Silver → Gold Medallion architecture in Databricks - ingesting raw CSV files from ADLS into Bronze, applying PySpark DataFrame transformations (deduplication, type casting, null handling, business-rule logic) in Silver, and producing aggregated, KPI-ready Gold tables – improving data reliability and reducing downstream query latency by ~35%.
- Leveraged Spark SQL and PySpark window functions for large-scale aggregations and incremental processing; applied partitioning strategies to reduce shuffle overhead cutting job execution time by ~25% - enabling the analytics team to access clean, structured datasets for reporting without manual intervention.
- Built an interactive sales analytics dashboard for business stakeholders with advanced DAX, dynamic KPIs and self-service drill-through.
- Designed a star schema data model and authored 25+ DAX measures including time-intelligence calculations delivering dynamic MoM/YoY KPIs that respond to any time granularity; reduced report query time by ~40% vs flat table predecessor through optimized model design.
- Built DAX-driven RAG Conditional Formatting, what-if scenario parameters and drill-through navigation - giving users instant visual health checks across 10+ KPIs, enabling them to interactively model revenue projections, and reach transaction-level detail in 2 clicks, reducing ad-hoc data requests to the analytics team by ~35%.
- RFM Scoring & Segmentation: Built a CTE-based RFM scoring model using window functions to segment the entire customer base into actionable tiers- Champions, Loyal, At-Risk, and Churned- delivering a reusable segmentation framework replacing ad-hoc manual analysis and cutting customer segmentation effort by ~60%, enabling the team to redirect that time towards strategy rather than data preparation.
- Churn Risk & Revenue Protection: Identified a high-value customer segment contributing ~30% of total revenue but showing early churn signals - enabling the commercial team to trigger targeted retention campaigns proactively.
- Developed a multi-view Tableau dashboard tracking supply chain KPIs using dual-axis charts, reference lines, and parameter-driven views to give users a real-time performance snapshot;
- Implemented complex business logic and transformation using Level of Detail (LOD) expressions, table calculations & parameters.
- Built dynamic parameter controls to switch between metrics and time periods within a single view; Connected to SQL-based data sources and configured Tableau Extracts with scheduled refresh - reducing dashboard load time by ~60%.
- Collaborated with stakeholders to translate operational reporting requirements into dashboard designs, iterating on layout, drill-down hierarchies, and tooltip detail to deliver a self-service tool that reduced manual reporting effort by ~30%.