Darshana Alone

Projects:

Project 1: Voice Bot

Description: Developed user-friendly voice bot to enhance customer experience on web & app. This bot leveraged LLM to understand user requests and provide natural conversation flow. It aimed to reduce call center cost.

Technologies Used: gpt-35-turbo, gpt-4, Microsoft Speech SDK, Open AI

Project 2: Consent-Based Gmail, SMS & Account Aggregator Data Model& Application

Description: This project aimed to improve apps like Due Bills and Money Manager. To do this, asked users for permission to access their Gmail and SMS data. With this data, used special computer programs and techniques (data science and language models) to uncover useful information. This information helped us make apps better and easier to use.

Technologies Used: Python, Azure App service, Functions, Cosmos DB, Blob Storage, LLM model, BERT

Project 3: SEO Nexus: Insights & Recommendation to drive more organic traffic

Description: Utilized SEO analytics, including keyword clustering, research, page rank correlation analysis, and content performance metrics to identify optimization opportunities. Implemented data-driven content optimization and generation, resulting in significant improvement in organic reach and user engagement. Integrated data sources used are GA, GSC, CRUX, Medalia, New Relic, Screaming Frog, Google KW planner, SEMrush

Technologies Used: Python, Clustering techniques, correlation analysis, Open AI, GPT, TF-IDF, Power BI

Project 4: Propensity Model

Description: Predict probability of customer to buy credit card based historical behavior, financial & demographics data.

Technologies Used: Python, Numpy, XGboost, Power BI,

Projects:

Project 1: Sales & Volume Forecasting Model

Description: Developed a cutting-edge Sales & Volume Forecasting Model for the Insurance and Retail industries. This dynamic model accurately predicts sales and volumes across segments, product lines, geographies, channels, and other key attributes. Applied analytics models, leveraging expertise in forecasting, resulting in optimized decision-making processes and enhanced business performance

Technologies Used : Python, Pandas, Numpy, Triple Exponential Smoothing, Arima, Prophet, Power BI, scikit-learn

Project 2: Return Propensity Scoring Model

Description: Created a precise Return Propensity Scoring Model for Retail. Predicts purchase return likelihood to optimize loyalty programs, enhance customer experience, and drive profitability.

Technologies Used: Python, Pandas, NumPy, Random Forest, Logistic Regression, Neural Net, Power BI

Projects:

Project 1: Sentiment Topic Analyzer

Description: Engineered a Customer Sentiment Topic Analyzer for Retail and NBFC. Employed web scraping for unstructured data extraction, utilizing advanced data science models for sentiment analysis and theme identification.

Technologies Used: Python, Pandas, LSTM, K-means, Genism, Selenium

Project 2: Product Sequencing & Recommendation Model

Description: This model prioritizes requested products based on customer behavior, profitability metrics, and company priorities. It integrates multiple parameters, including product profitability, customer demographics, similar customer preferences, customer value score, past orders, seasonality, and profitable product combinations. The model enhances decision-making by providing a strategic approach to product sequencing and recommendations, optimizing both customer satisfaction and company profitability.

Technologies Used: Python, Pandas, Numpy, Apriori Algorithm, POwer BI, k -means

Projects:

Project 1: Web application & Rest API Development

Technologies Used: Python, Django