Swanand Marathe
Business Intelligence Analyst
Hello
Powering Product Growth Through Data, AI, and Strategy
My journey began in on-ground sales, where I gained a firsthand understanding of user behaviour and system friction, shaping a product mindset rooted in real-world problems, not just feature delivery.
Building on this foundation, I’ve worked across Business Intelligence, Product Analytics, and Marketing — supported by a Master’s in Business Management and certifications in Product Management, Generative AI, and Business Analytics. Today, I build GenAI- and ML-powered products at the intersection of data, AI, and user experience.
From AI agents for EHR automation to prompt pipelines and Mixpanel-driven insights, my focus is on scalable, intelligent systems that automate, inform, and elevate product decision-making.
Projects
Please select the domain from the Menu.
Generative AI
-
Purpose: Designing a GenAI-powered chatbot agent to manage complex, multi-step medical interactions such as providing medical information, locating nearby doctors, and facilitating healthcare utilities through a single conversational interface.
-
Platform and Process: Built using a modular orchestration framework to manage tool-based workflows, a cloud-based LLM deployment environment, and an integrated vector database for retrieval-augmented generation (RAG). The system leveraged agentic reasoning and task chaining architectures to enable dynamic decision-making and maintain contextual flow across healthcare-specific interactions.
-
Skills Utilized: Applied expertise in generative AI, agent design, RAG pipelines, conversational flow management, and orchestrated execution of multi-component systems — all tailored to the complexity and compliance needs of healthcare applications.
EHR Automation from Unstructured Medical Documents
-
Purpose: Developed a GenAI-driven workflow to transform unstructured medical documents into structured, clinically aligned EHR outputs for improved healthcare data accessibility and integration.
-
Platform and Process: Implemented using a context-aware generation framework incorporating retrieval-augmented generation (RAG) via vector-based semantic search and document chaining techniques. Ensured adherence to clinical ontologies (FHIR, SNOMED CT, LOINC) while leveraging AI workflow orchestration layers to enhance scalability, accuracy, and compatibility with real-world healthcare systems.
-
Skills Utilized: Showcased capabilities in clinical NLP, GenAI-based document structuring, semantic search integration, standards-driven data modeling (FHIR, SNOMED CT), and deploying orchestrated AI pipelines within a healthcare data processing environment.
Machine Learning
Please click on button to explore the projects in Github.
.png)
Building Churn Prediction Model Using Jupyter Notebook
-
Purpose: Develop a churn prediction model using Jupyter Notebook to analyze and predict customer churn by employing various data analysis and machine learning techniques.
-
Platform and Process: Utilized Jupyter Notebook for a step-by-step analysis, including data loading, exploratory data analysis (EDA) techniques such as univariate, bivariate, and multivariate analysis, data cleaning, principal component analysis (PCA), clustering, and model building using algorithms like KNN, Random Forest, SVM, ADABoost, and XGBoost.
-
Skills Utilized: Demonstrated proficiency in data analysis, machine learning, and Python programming within Jupyter Notebook, showcasing the ability to develop and evaluate complex predictive models for churn prediction.
.png)
Time Series Forecasting Using Jupyter Notebook
-
Purpose: Utilizing Jupyter Notebook for time series forecasting to predict shoe company and soft drink sales, aiming to refine predictions and enhance precision in predictive analytics using various models.
-
Platform and Process: Employed Jupyter Notebook for a methodological progression in sales forecasting, starting from Linear Regression and Naive Forecast for shoe company sales, and advancing to the Triple Exponential Model. For soft drink sales, the project utilized the Seasonal Autoregressive Integrated Moving Average (SARIMA) model.
-
Skills Utilized: Demonstrated proficiency in time series forecasting techniques, model selection, and refinement using Jupyter Notebook, showcasing the ability to develop accurate predictive models for sales forecasting.
Product and Brand
Please click on button to open project document in pdf.
Marketing Plan for Brandy company Entering Spanish Market
-
Purpose: Developed a marketing plan for CR&F, a brandy company, to enter the Spanish market strategically, aiming to establish a strong brand presence and capture market share.
-
Platform and Process: Utilized SPSS for customer analysis and segmentation based on survey data interpretation, providing insights into customer preferences and behaviour. Conducted thorough market and competitor analyses to inform strategic decision-making and devised a comprehensive brand-building strategy with targeted campaigns.
-
Skills Utilized: Demonstrated proficiency in marketing strategy development, customer analysis, survey data interpretation using SPSS, market analysis, and campaign planning, showcasing the ability to devise effective strategies for entering new markets.
.png)
CPG Brand Creation and Management
-
Purpose: Introducing "Protein Flakes" in the Consumer Packaged Goods (CPG) industry through a comprehensive brand creation and management strategy to establish market presence and drive consumer engagement.
-
Platform and Process: Utilized Wix to build a website as part of the brand management strategy. Employed various brand management tools including value proposition, brand personality, buyer personas, customer journey mapping, and digital marketing techniques.
-
Skills Utilized: Demonstrated proficiency in brand creation and management, utilizing tools such as value proposition, brand personality development, buyer persona creation, customer journey mapping, and digital marketing strategies to effectively launch and manage the "Protein Flakes" brand within the CPG industry.
Data Analytics
Please click on button to explore the projects in Github.
.png)
-
Purpose: Utilizing BigQuery to analyze the annual sales data of a bike store to generate a comprehensive Sales Report.
-
Platform and Process: Employed SQL querying techniques on BigQuery platform to extract and manipulate data for in-depth analysis.
-
Skills Utilized: Demonstrated proficiency in advanced SQL querying, data extraction, and proficient database management for accurate interpretation of sales data.
.png)
Data Mining (PCA and Clustering) using
Jupyter Notebook
-
Purpose: Utilizing Jupyter Notebook for implementing Principal Component Analysis (PCA) on customer data and clustering algorithms on health-income data to extract meaningful patterns and insights.
-
Platform and Process: Implemented PCA and clustering algorithms using Python within the Jupyter Notebook environment for comprehensive data mining and analysis.
-
Skills Utilized: Demonstrated proficiency in PCA, data clustering techniques, and Python programming, showcasing advanced capabilities in data analysis methods and tools.
.png)
RFM Segmentation and
Market Basket Analysis using KNIME
-
Purpose: Employing KNIME for RFM segmentation on sales data and performing market basket analysis using association rules on retail store data to gain insights into customer behaviour and optimize marketing strategies.
-
Platform and Process: Utilized KNIME workflows to execute RFM segmentation and association rule-based techniques for market basket analysis, leveraging its capabilities in data processing and analytics.
-
Skills Utilized: Demonstrated proficiency in KNIME, RFM analysis, association rule-based techniques, and retail data analytics, showcasing expertise in leveraging tools and methodologies for customer segmentation and market analysis.
Data Visualisation
Please click on the button to explore dashboards in Tableau Public.
.png)
Election Voter Turnout and Results Analysis Using Tableau
-
Purpose: Utilizing Tableau for in-depth analysis of voter turnout and election results in Maharashtra during the 2014 and 2019 general elections, aiming to understand electoral dynamics and political shifts.
-
Platform and Process: Leveraged advanced Tableau techniques such as dynamic parameters, calculated fields, and sophisticated filters to visualize and analyze the electoral landscape effectively.
-
Skills Utilized: Demonstrated proficiency in Tableau visualization, data analysis, and understanding of electoral processes, showcasing the ability to derive nuanced insights from complex datasets for informed decision-making.
.png)
IOWA Liquor Sales Analysis for 2021 Using Tableau
-
Purpose: Creating a detailed sales analysis report for Iowa Liquor sales in 2021 using Tableau, integrating spatial data for mapping and government election information to provide a holistic view of sales trends and patterns.
-
Platform and Process: Utilized Tableau to develop the sales analysis report, incorporating spatial data and government election information for comprehensive visualization and analysis of liquor sales data.
-
Skills Utilized: Demonstrated proficiency in data visualization using Tableau and the integration of diverse datasets, showcasing the ability to analyze and present complex data effectively for informed decision-making.