Shubham Jaiswal

Data Analyst | SQL · Python · Power BI · DAX | Turning Raw Data Into Revenue Decisions | Unit Economics · Churn Analysis · BI Analytics | Open to Remote Roles

[email protected] +91 8081008272

Resume

Hello, I am

Shubham Jaiswal.

4

PowerBI
Projects

2

SQL
Projects

2

Python
Projects

About Me

I turn raw data into revenue decisions. Before data analytics, I spent 3 years running sales operations at a building materials business — managing pricing, tracking margins, and watching inventory move manually. That experience taught me what a business problem feels like before I open a query editor.

In 2025 I pivoted into data analytics. Since then I have built a Power BI dashboard processing 1.2M+ insurance policy records for C-suite decision making, identified post-session payment friction via SQL behavioral analysis that pivoted a failing product to a subscription model,
and modeled a 14.7x LTV:CAC ratio live under executive scrutiny — earning a Letter of Recommendation from the CSO.

My toolkit covers Advanced SQL (CTEs, Window Functions), Python (Pandas, NumPy), Power BI (DAX, Star Schema), and Databricks.

I am actively looking for remote Data Analyst or BI Analyst roles where the work actually moves the business.

Key Skills

Excel

Power Bi

Python

SQL

Databricks

Data Modeling

Business Strategy

Unit Economics

My Projects

Production Insurance Analytics: 1.2M+ Policy Records Performance Dashboard

Domain/Function: Insurance / Sales Analytics

Decoding the 2026 Tamil Nadu Election | AtliQ Media Pitch

Domain/Function: Media & Political Data Analytics

Business Insights 360 Dashboard for Finance, Sales, Marketing and Supply Chain

Domain/Function: Electronics

Feelio: AI-Driven Mental Health & Therapy Platform

Domain/Function: Healthcare / Mental Wellness & Product Architecture

E-Commerce Churn Lifecycle: End-to-End Enterprise AI Solution

Domain/Function: E-commerce

Customer Churn & Revenue Leakage Analysis — E-Commerce Platform

Domain/Function: E-Commerce / Retail

Wavecon Telecom: 5G Launch Impact & Revenue Paradox Analysis

Domain/Function: Telecom / Revenue Assurance

SQL Detective: The Six-Week Reality Restoration Project

Domain/Function: Data Forensic / Investigation

SQL Murder Mystery: The TechNova CEO Investigation

Domain/Function: Data Forensics / Logic & Deduction

My Experience

Data Analyst | Internships & Business Operations

AtliQ Technologies (Dec 2025 – Jan 2026): Built a Power BI dashboard processing 1.2M+ insurance policy records with Star Schema modeling and MoM trend analysis for C-suite decisions. Analyzed Wavecon Telecom 5G launch impact and automated weekly reporting pipelines using Power Query.
Digits n Data (Feb 2026 – Mar 2026): Modeled a 14.7x LTV:CAC ratio for Feelio therapy app. Led 6-member cross-functional team through live executive pitch. Earned CSO Letter of Recommendation for strategic business defense.
Prakash Building Material Store (Aug 2022 – Sep 2025): Managed pricing, margins, supplier negotiations and inventory for a building materials business — building real business context before entering data analytics in 2025.

Awards & Certificate

45-Day Internship Program (Data Analysis & Business Strategy)

Data Analytics Bootcamp 5.0: Job Placement Support + AI Automation & Data Engineering Basics

Databricks 14 Days AI Challenge

Let's Connect

Feel free to get in touch with me. I am always open to discussing new projects, creative ideas or opportunities to be part of your visions.

Mail Me

[email protected]

https://www.linkedin.com/in/iam-shubhamjaiswal/

Call Me

+91 8081008272

GitHub

https://github.com/JaiswalAnalytics

PowerBI Project

Production Insurance Analytics: 1.2M+ Policy Records Performance Dashboard

Domain/Function: Insurance / Sales Analytics

Developed a production-ready Power BI dashboard analyzing 1.2M+ insurance policy records. Implemented a Star Schema to track Revenue growth, MoM policy trends, and settlement risk. Delivered actionable insights on sales channel performance and demographic revenue distribution.

View Project
Engagement

View Project
on GitHUB

Other Tools

DAX

Power Query

Data Modeling (Star Schema)

Excel

Comprehensive Insurance Executive Dashboard featuring high-level KPIs, Revenue vs. Customer growth trends, and sales channel distribution analysis for 1.2M records.

Data Architecture view showing the Star Schema implementation with optimized relationships between dimension and fact tables to support high-speed analytical queries.

Project Details

As part of the Codebasics Virtual Internship, I engineered an end-to-end insurance analytics solution designed for executive-level decision-making.

Technical Core:
• Data Modeling: Architected a robust Star Schema (dim_customer to fact_premiums) to ensure high-performance filtering across 1.2M+ rows.
• Advanced DAX: Authored complex measures for Day-over-Day (DoD) growth, Month-over-Month (MoM) trends, and a dynamic trend toggle allowing users to switch between Revenue and Customer views instantly.
• Insight Engineering: Created age-banding logic and city-wise matrices to identify that the 31-40 age group is the primary revenue driver.
• Sales Strategy: Analyzed channel efficiency, revealing the Offline-Agent channel's dominance at 56.67% of total sales.

The final dashboard provides a comprehensive view of settlement risks and growth trajectories, enabling stakeholders to optimize marketing spend and agent resources based on real-time data performance.

Project video

Live Interactive Dashboard

Python Project

Decoding the 2026 Tamil Nadu Election | AtliQ Media Pitch

Domain/Function: Media & Political Data Analytics

Analyzed all 234 constituencies from the 2026 Tamil Nadu Assembly election using Python and ECI data. Built 3 data-driven TV stories and a 10-slide stakeholder pitch deck for fictional news network AtliQ Media.

View Project
Engagement

View Project
on GitHUB

View Project
on Codebasics Resume Project Challenge

Other Tools

Pandas

Matplotlib

Canva

Jupyter Notebook

ECI Data

Regional flip rate breakdown across Tamil Nadu's 6 editorial regions. Chennai Metro led at 94% while Delta was most stable at 45.5%. Built using Python and Matplotlib, visualized in the Canva pitch deck.

Story 3 key stats — average winning margin collapsed from 11.7% to 7.7%, 64 winners got under 35% vote share, and one constituency was decided by exactly 1 vote in 2026.

Project Details

AtliQ Media wanted a clean, fact-based one-hour TV show on the 2026 Tamil Nadu Assembly election results with no political commentary — only ECI data turned into stories.

I analyzed 4,257 candidate-level rows across 234 constituencies, cleaned and joined 2021 and 2026 datasets using ac_number as the primary key, and built a master analysis table in Python.

Three stories were selected and verified against raw data:

Story 1 — The Map Redrawn: 163 of 234 seats changed party in one election. TVK won 108 seats in its very first election.

Story 2 — The Wave Hit Unevenly: Chennai Metro saw 94% of seats flip. Delta was the most stable at 45%.

Story 3 — The Illusion of Decisive Wins: 64 winners got less than 35% of votes. Average margins collapsed from 11.7% to 7.7%. One seat was won by exactly 1 vote.

Deliverables: 5 Python charts, 10-slide Canva pitch deck, professional GitHub repo with documented notebooks.

Tools: Python, Pandas, Matplotlib, Canva, Jupyter Notebook.

PowerBI Project

Business Insights 360 Dashboard for Finance, Sales, Marketing and Supply Chain

Domain/Function: Electronics

AtliQ Hardwares is a consumer electronics company expanding rapidly but is not able to compete with other companies using data as most of their report still exists in Excel. My goal is to implement an advanced analytics solution using Power BI that will enable the company to get insights and make in

Other Tools

Excel

SQL

Power Bi

User-empathetic landing page as a part of dashboard design

Data modelling involving 10+ tables

Project Details

In this project, the goal is to be one report which could be used by stakeholders from sales, marketing, finance and executive team. The focus is made on the following

1. Robust Data Modeling
2. User-empathetic Report design
3. Drillable Insights

I Explored the following things in this project:

1. Power Query (Basic and Advanced Operations)
2. Basic and complex DAX formulas
3. Data modelling involving 10+ tables
4. Choosing the right visuals and formatting
5. Dashboard designing principles
6. Using bookmarks
7. Deploying in Power BI service
8. Stakeholder Feedback Implementation
9. Sales, Marketing, Finance & Supply Chain metrics

Live Interactive Dashboard

Feelio: AI-Driven Mental Health & Therapy Platform

Domain/Function: Healthcare / Mental Wellness & Product Architecture

Feelio is a comprehensive therapy app designed to bridge the gap between mental health professionals and users. It features mood tracking, session scheduling, and AI-driven insights to improve emotional well-being.

View Project
Engagement

Other Tools

SQL

Excel

The Feelio logo and tagline represent the core mission of the platform: providing a safe, supportive bridge between individuals and mental wellness.

Letter of Recommendation: CSO, Digits n Data Official commendation recognizing my collaborative leadership, logical reasoning, and ability to defend strategic business decisions during live executive presentations.

Project Details

Feelio is an on-demand therapy platform rebuilt through a data-driven pivot. Facing ₹24,200/month burn, 14.1% churn, and 4.5 months of runway, we restructured core business logic for survival and scale.
Key Interventions:
Model Pivot — Replaced Pay-Per-Session with a ₹499/month Care Plan, shifting from transactional fixes to continuous "Access Value."
Friction Reduction — SQL analysis revealed 45.5% of churn came from payment barriers. We streamlined checkout to support users at their most vulnerable.
Retention Engine — Targeted 29.3% "post-session silence" churn by building daily habit loops, transforming Feelio from one-off utility to wellness companion.
Human Moat — Positioned as the "Fast + Human" alternative to AI bots and long waitlists.
By converting 23% of active users, we optimized LTV:CAC and built a sustainable unit economics engine.

Python Project

E-Commerce Churn Lifecycle: End-to-End Enterprise AI Solution

Domain/Function: E-commerce

Built a robust Data & AI solution on Databricks to predict customer churn. Implemented a Medallion Pipeline with Unity Catalog governance and MLflow tracking. Identified 5,881 high-risk users, automating a 25% discount strategy to protect revenue.

View Project
Engagement

View Project
on GitHUB

Other Tools

Databricks

MLflow

Unity Catalog

PySpark

Python

Medallion Architecture pipeline in Databricks showing the transformation from raw ingestion to Gold-level business aggregates.

MLflow tracking UI displaying model parameters, metrics, and Run ID 9b9d1... for experiment reproducibility.

Project Details

This project, completed as part of the 14-day Databricks challenge, demonstrates a production-grade AI lifecycle for e-commerce retention.

Key Technical Implementation:
• Data Engineering: Engineered a Medallion Architecture (Bronze-Silver-Gold) using PySpark to transform raw customer behavior data into analysis-ready gold tables.
• Governance: Leveraged Unity Catalog for fine-grained access control and full data lineage, ensuring enterprise-level security.
• Machine Learning & MLOps: Developed a churn prediction model. Used MLflow for experiment tracking (Run ID: 9b9d1...), ensuring reproducibility and model versioning.
• Business Logic: Segmented 5,881 high-risk customers and designed an automated retention workflow offering 25% discounts.

This solution moves beyond simple modeling into a governed, scalable architecture, bridging the gap between Data Engineering and Data Science to drive measurable business impact.

Project video

PowerBI Project

Customer Churn & Revenue Leakage Analysis — E-Commerce Platform

Domain/Function: E-Commerce / Retail

Analysed 99,441 orders & 96,096 customers from Olist e-commerce using Python, SQL & Power BI to identify R$6.45M (40.1%) revenue at risk from churning customers via RFM segmentation & cohort retention analysis.

View Project
Engagement

View Project
on GitHUB

Other Tools

Excel

Pandas

Matplotlib

Seaborn

Executive Summary page — 4 KPI cards (Total Revenue R$16.1M, Churn Rate 39.99%, Revenue at Risk R$6.45M, Recoverable R$3.7M), customer segment bar chart, and revenue donut chart by segment.

RFM Analysis page — revenue by segment bar chart (At Risk holds R$3.9M), behavioural profile matrix with avg recency/frequency/monetary, Recency vs Monetary scatter plot, and interactive segment slicer.

Project Details

Analysed 99,441 orders & 96,096 customers from Olist e-commerce to identify R$6.45M (40.1%) revenue at risk from churning customers.

Approach:
- Merged 9 datasets into a master table using Python & SQL
- Built RFM segmentation — 96,096 customers across 5 behavioural tiers
- Cohort retention revealed <1% Month-2 retention across all cohorts
- Delivery delay >4 days flagged as primary churn driver (score drops 4.03 → 2.10)
- 22,975 Potential Loyalists = R$3.7M recoverable with one campaign

Key Findings:
- At Risk segment holds R$3.9M — more than Champions (R$2.8M)
- Even Champions average only 1.11 purchases — a frequency problem, not spend
- 47.8% of customers are At Risk or Lost

Deliverable: 3-page interactive Power BI dashboard with DAX measures tracking Churn Rate, Revenue at Risk & Cohort Retention, plus 3 data-backed business recommendations.

Live Interactive Dashboard

PowerBI Project

Wavecon Telecom: 5G Launch Impact & Revenue Paradox Analysis

Domain/Function: Telecom / Revenue Assurance

A strategic Power BI analysis of a 5G network rollout. Identified a 'Revenue Paradox' where market share grew to 20.97% while total revenue declined by 0.50%. Provided geo-specific recommendations and product-level insights to stabilize Average Revenue Per User (ARPU).

View Project
Engagement

Other Tools

Advanced DAX

Competitive Benchmarking

Market Share Analysis

Comparative analysis dashboard visualizing the 20.97% market share growth against the -0.50% revenue decline, highlighting the 'Revenue Paradox' post-5G launch.

Geospatial performance analysis showcasing localized revenue shifts, with a focus on Tier 2 growth (Lucknow) and critical Metro market losses.

Project Details

This project evaluates Wavecon Telecom's business performance following its 5G launch. Using Power BI, I analyzed high-volume data to uncover a "Revenue Paradox": market share grew to 20.97%, yet total revenue declined by 0.50%.

Technical & Analytical Highlights:
• KPI Analysis: Identified a disconnect between user acquisition and profitability, signaling aggressive pricing or high-value subscriber churn.
• Geographic Insights: Discovered Tier 2 city growth (Lucknow +1.82%) vs. major metro decline (Chennai/Delhi), suggesting a need for regionalized strategies.
• Product Impact: Isolated the 'Smart Recharge Combo' as a top performer (+₹0.6 Bn) and audited the risks of sunsetting Plan P13.
• DAX & Modeling: Built dynamic measures for Pre-vs-Post 5G comparisons and competitive benchmarking against rivals.

This analysis provides a roadmap for stabilizing ARPU and optimizing 5G migration to convert market share into sustainable revenue.

Project video

SQL Project

SQL Detective: The Six-Week Reality Restoration Project

Domain/Function: Data Forensic / Investigation

A six-week intensive SQL investigation solving complex crises through data forensics. Leveraged advanced querying—from subqueries to window functions—to identify security breaches, track $1.6M in illicit transfers, and predict temporal attack sequences.

View Project
Engagement

Other Tools

PostgreSQL (or MySQL)

Advanced Window Functions

Data Cleaning

Detailed view of a forensic SQL investigation utilizing JOINs, GROUP BY, and Date Filtering to track $1.6M in illicit transfers. Includes the specific query deployed to aggregate transaction data and identify the 'Vision AI Research' stronghold.

A comprehensive technical roadmap of the six-week project. It showcases a diverse SQL toolkit, including Window Functions (LEAD), Subqueries for benchmarking, and multi-column Advanced Sorting used to solve complex, logic-driven data challenges.

Project Details

Successfully completed the "Six Weeks to Save Reality" challenge, acting as a Data Analyst Detective to resolve six critical timeline crises using PostgreSQL. This project showcases the transition from basic syntax to high-level analytical logic.

Technical Milestones:
• Benchmarking & Subqueries: Dynamically identified top-performing responders based on real-time energy spikes.
• Geographic Instability Analysis: Used sorting and limits to prioritize global deployment zones based on stability indices.
• Forensic Filtering: Conducted 48-hour window analysis and date filtering to expose internal security overrides.
• Financial Auditing: Aggregated $1.6M in black-market transfers using GROUP BY and SUM to pinpoint coalition strongholds.
• Predictive Modeling with LEAD(): Implemented Window Functions to sequence time-series data and predict attack timestamps with second-level precision.
• Multi-Layered Logic: Used complex tiebreaker sorting for optimal landing point selection.

SQL Project

SQL Murder Mystery: The TechNova CEO Investigation

Domain/Function: Data Forensics / Logic & Deduction

A data-driven murder mystery investigation solved using advanced SQL. Acted as a Lead Analyst to cross-reference keycard logs, phone records, and alibis through complex JOINs and CTEs to identify the culprit behind a corporate homicide case.

View Project
Engagement

Other Tools

PostgreSQL/MySQL

CTEs

Multi-table JOINs

Subqueries

Advanced SQL query using CTEs and multiple JOINs to aggregate suspect movement, communication logs, and physical evidence for the final deduction.

Visual breakdown of the investigative roadmap, highlighting the data-driven steps taken to verify alibis and isolate the prime suspect.

Project Details

In this capstone project, I utilized SQL as a forensic tool to solve a simulated murder mystery at TechNova Inc. The project required deep logical deduction and complex querying to turn raw evidence into a final conviction.

Technical Highlights:
• Temporal Analysis: Used WHERE clauses and BETWEEN filters to isolate employee movements within a 30-minute crime window (8:45-9:15 PM).
• Forensic Joining: Implemented multi-table JOINs on keycard logs and employee directories to identify high-suspicion individuals.
• Alibi Verification: Leveraged LEFT JOINs and Subqueries to cross-reference suspect statements against digital footprints, exposing false claims.
• Advanced Logic: Built Common Table Expressions (CTEs) to aggregate communication logs and physical evidence into a single investigative view.

This project sharpened my ability to perform complex data auditing and storytelling, essential skills for Business Intelligence and Fraud Detection roles.