Niharika Kotta

Net Spam a Network-based Spam Detection

• Many people rely on social media reviews for decisions, which attracts spammers who post fake feedback about products. NetSpam is a new method that identifies these spammers and fake reviews by analyzing review data as a complex network.This approach turns the challenge of spam detection into an easier classification problem within that network.
  
  Caption generator using Machine Learning
• CNNs (from deep learning) first analyze the image to identify key visual features like objects and their relationships.The extracted features are then passed to LSTMs (a type of RNN/NLP) to sequentially generate words based on language rules.The entire system is trained on vast datasets of paired images and captions, allowing it to connect the visual world with descriptive text.

Data Analyst Internship: Leveraged SQL to query, clean, and manipulate large datasets. Developed and deployed interactive dashboards in Tableau and Power BI, translating complex analyses into actionable business insights that contributed to a key operational decision and improved reporting efficiency. Demonstrated proficiency across the entire data lifecycle, from extraction and transformation to visualization and strategic communication.

Zomato Analysis

Used SQL to clean and prepare a large Zomato dataset, followed by Excel modeling to calculate key performance indicators (KPIs) and understand consumer preferences.Designed interactive dashboards in both Tableau and Power BI to visualize the findings.The project successfully revealed critical trends in restaurant ratings and cuisine popularity, which were used to inform and shape market strategy.

Adventure Works

This project focused on converting raw, unstructured data from the Adventure Works database into a powerful analytical model that supports strategic business insights. calculating crucial financial metrics like Sales Amount, Production Cost, and Profit.