I am a business professional with experience in Business Development, Sales, CRM, Manufacturing and Entrepreneurship, now specializing in Data Analytics, Business Intelligence and data-driven decision-making.
I combine business understanding with data and technology to transform business problems into measurable insights, actionable recommendations and business decisions.
Business Problem → Data → Analysis → Insight → Action
📍 Bangalore, India
💼 Assistant Manager – Business Development
🎓 Pursuing MS in Data Science through the Scaler–Woolf pathway
🎯 Target roles: Data Analyst | Business Analyst | BI Analyst
- 🌐 Portfolio: vijaybsbs.github.io
- 💼 LinkedIn: Vijay Kumar
- 💻 GitHub: @vijaybsbs
BigQuery / GoogleSQL / Looker Studio
End-to-end e-commerce analytics covering data quality, customer behaviour, geography, order economics, payments, logistics, delivery performance, customer experience and RFM analysis.
Selected findings & business implications
- 97% of active customers are one-time buyers; repeat customers represent only 3% → retention initiatives should focus on converting one-time buyers into repeat customers.
- 48.85% of customers account for 80% of observed customer value → customer-value concentration can help prioritize retention and targeted engagement.
- São Paulo, Rio de Janeiro and Minas Gerais represent 61.24% of recent high-value one-time customers → these markets provide a clear geographic focus for retention and cross-sell initiatives.
- 80.82% of recent high-value one-time orders contain a single item → cross-sell and basket-expansion opportunities should be investigated for high-value one-time customers.
Key skills: SQL • BigQuery • Data Modelling • CTEs • Window Functions • RFM • KPI Analysis • Dashboarding
BigQuery / GoogleSQL / Looker Studio
Retail sales and inventory analysis across 50 stores and 35 products, covering revenue, cost, product performance, store performance, pricing, inventory risk and Pareto analysis.
Selected findings & business implications
- $14.44M revenue, $4.01M gross profit and 27.79% gross margin.
- The top 15 of 35 products (43%) generate approximately 80.08% of revenue → revenue is concentrated, but not at a classic 80/20 level; availability and replenishment of these high-contribution products deserve priority.
- Toys is the largest revenue category at 35.26%, while Electronics has the highest gross margin at 44.57% → category strategy should balance revenue scale with margin contribution.
- 157 store-product combinations have no inventory record, while 7 duplicate inventory rows require data-quality attention → inventory decisions should account for incomplete and duplicated inventory records.
- 3 inventory records have stock despite no historical sales → these combinations warrant review for assortment fit, local demand or potential overstock.
Key skills: SQL • Window Functions • Ranking • Pareto Analysis • Retail Analytics • Inventory Analysis • Dashboarding
Python / EDA / Statistics
Bike rental demand analysis using exploratory data analysis and statistical hypothesis testing to evaluate the impact of working days, seasons, weather and environmental factors on demand.
Statistical findings
- Working days: independent 2-sample t-test → t = 1.210, p = 0.2264 → no statistically significant difference in average rentals between working and non-working days.
- Season: one-way ANOVA → F = 236.95, p = 6.16 × 10⁻¹⁴⁹ → rental demand differs significantly across seasons; Fall has the highest median and Spring the lowest in the analysis.
- Weather: one-way ANOVA → F = 65.53, p = 5.48 × 10⁻⁴² → rental demand differs significantly across weather conditions; clear weather has the highest median, while light snow has the lowest.
- Season × Weather: chi-square test → χ² = 46.10, df = 6, p = 2.83 × 10⁻⁸ after excluding the single heavy-rain observation because the original expected-frequency assumption was violated → season and weather are statistically associated.
Business implication: demand planning should account for season and weather conditions rather than treating working-day status alone as a significant demand driver.
Key skills: Python • Pandas • EDA • Statistics • Hypothesis Testing • Business Analytics
Tableau
Interactive business intelligence dashboard focused on sales performance, revenue targets, regional performance, freight costs, product-segment trends and customer retention.
Selected findings & business implications
- Revenue reached 9.24M against a 9.751M target, representing approximately 94.8% target achievement → the business was below its revenue target by about 0.51M in the dashboard period.
- Order quantity was 129,284 against a target of 372,420, representing approximately 34.7% target achievement → quantity performance shows a substantial gap relative to the target.
- The dashboard reports an average discount of 4.96%, freight expense of 119,699 and an average unit price of 88, providing a basis for monitoring pricing, discounting and logistics costs together.
- Regional revenue varies materially across the states shown in the dashboard, including South Australia (962,572), Tasmania (795,403) and Northern Territory (514,268) → regional performance can be compared to identify differences in revenue contribution.
- Revenue trends vary across wine segments over time: Red Wine shows a relatively consistent pattern, while Rose/Sparkling Wine shows greater volatility and White Wine records a pronounced revenue spike around 2012.
- Order quantity increases across the displayed years, rising from approximately 25K in 2010 to 42K in 2013 → the dashboard indicates increasing order volume over the period.
- The customer retention view shows a growing contribution from customers associated with earlier ordering periods over time, providing a basis for monitoring repeat-order behaviour and customer retention.
Business focus: revenue performance • target tracking • regional analysis • wine-segment trends • freight analysis • customer retention
SQL + Python / Pandas + Power BI / DAX
An end-to-end restaurant analytics case study covering SQL business questions, exploratory data analysis, data-quality validation, Python-based cleaning and feature engineering, and a six-page Power BI dashboard.
The project evolved through three phases: SQL analysis → Python EDA → corrected analysis and interactive BI reporting. Deeper investigation uncovered rating placeholders and sampling limitations that changed several earlier conclusions.
Selected findings & business implications
- 90.59% of the original 9,551 records are from India (8,652 restaurants). Because the dataset mixes countries and contains inconsistent currency labels, the final pricing and market analysis focuses on India.
- Data quality changed the rating story: 2,139 Indian restaurants with fewer than four votes had a placeholder rating of 1.0. Treating these as not rated yet raised the average Indian rating from 2.77 to 3.35.
- City comparisons required a sampling check: 28 Indian cities had exactly 20 listings. The final analysis restricts city-level comparisons to cities with at least 100 restaurants, avoiding misleading conclusions from small samples.
- Delhi NCR accounts for 92% of India's listed restaurants (7,947 of 8,652), with New Delhi alone contributing 63%—highlighting the dataset's strong geographic concentration.
- Price and ratings: cleaned average ratings rise from 3.20 in Budget to 3.74 in Luxury, but the difference between Premium and Luxury is only 0.06 rating points, despite roughly twice the cost.
- Delivery opportunity: 64% of restaurants offer neither online delivery nor table booking. Gurgaon leads the qualifying city comparisons on delivery adoption at 38.0%, while New Delhi's scale suggests a substantial potential delivery gap.
Power BI dashboard: Six pages covering dataset scope, data quality, market overview, ratings and engagement, price and value, and delivery and services.
Key skills: SQL • Python • Pandas • Data Cleaning • Feature Engineering • EDA • Power BI • DAX • Data Modelling • Business Analytics
📄 View 6-Page Power BI Dashboard (PDF)
Python / Pandas
Logistics data analysis and feature engineering to transform trip-level operational data into model-ready features for route, delivery and performance analysis.
Selected findings & business implications
- 144,867 operational records represent 14,817 unique trips across approximately 27 days of data.
- Average actual trip time is 417 minutes vs 214 minutes estimated by OSRM → actual operations take substantially longer than routing estimates on average, making estimate-vs-actual gaps important for operational planning.
- Average actual distance is 234 km vs 285 km OSRM estimated distance → routing distance and observed operational distance differ materially and should be evaluated separately rather than treated as interchangeable measures.
- The dataset contains 1,500+ source/destination logistics centers, providing a broad geographic basis for route and hub-level analysis.
- 293 source-name and 261 destination-name values were missing before preprocessing, representing less than 0.3% of the 144,867 records; no duplicate records were identified.
Business focus: trip-level grain • route performance • actual vs estimated distance/time • operational feature engineering
Key skills: Python • Pandas • Data Cleaning • Feature Engineering • Operational Analytics
Sales • Revenue Analysis • CRM Analytics • Customer Analytics • Lead Funnel Analysis • Market Analysis • KPI Reporting
SQL • GoogleSQL • BigQuery • Python • Pandas • Excel • Exploratory Data Analysis • Statistics
Tableau • Looker Studio • Dashboarding • KPI Analysis • Data Visualization
Generative AI • Prompt Engineering • AI-assisted Analytics
Git • GitHub • Google Colab • Jupyter • Google Sheets
My career sits at the intersection of business, sales, data and technology.
I bring practical experience across Business Development, Sales, CRM, Revenue Analysis, Manufacturing and Entrepreneurship, with a growing focus on data-driven decision-making.
- Sales and revenue performance analysis; CRM activity and lead-funnel analytics
- Daily sales productivity, KPI tracking and performance dashboards
- Created a lead generation → nurture → conversion SOP supported by daily sales reporting
- Architect, builder and dealer/channel engagement; project and specification business
- Cross-functional coordination with marketing and sales teams
- Conducted knowledge-sharing / L&L sessions for design and execution teams
I use SQL, Python, Excel, BigQuery and BI tools to connect business questions with data and translate analysis into actionable business insights.
Business Problem → Data → Analysis → Insight → Action
I am pursuing a Master of Science (MS) in Data Science through the Scaler learning pathway, with the academic degree awarded by Woolf Higher Education Institution.
Core Areas:
Data Analytics • Statistics • SQL • Python • Machine Learning • Data Visualisation • Business Intelligence • Applied Data Science
Saïd Business School, University of Oxford
Completed the Introduction to Advanced Business Analytics with AI course, focused on applying business analytics and AI to data-driven decision-making, strategic and operational outcomes, and responsible AI adoption.
Key Areas:
Business Analytics • Generative AI • AI for Decision-Making • Predictive & Prescriptive Analytics • AI Risk & Ethics • Real-World Case Studies
- Completed: August 2026
- Academic Credit: 1 credit
- Credential: University of Oxford, Saïd Business School
Business Management • Strategy • Marketing • Operations • Entrepreneurship
- Advanced SQL and analytical problem solving
- Python for data science and applied analytics
- Statistics and applied data analysis
- Generative AI and LLM applications
- AI-assisted business analytics
I am interested in opportunities where I can combine business understanding, analytics and technology to solve real-world business problems.
Primary: Data Analyst • Business Analyst • BI Analyst
Growth Areas: Business Analytics • Applied Analytics • Data Science • AI & Analytics