Surya Vaddhiparthy's Projects
This Python code analyzes a dataset of blog posts, focusing on the polarity and subjectivity of the text. It cleans the text data, visualizes word frequency using word clouds, and explores the sentiment of the text based on age groups and blog topics. The results show differences in sentiment, subjectivity, and word usage among age groups and blog
Leveraging advanced data analytics methodologies and time series forecasting with ARIMA modeling, this project delivers a comprehensive analysis of COVID-19 trends and metrics in the United States, providing crucial insights for informed decision-making in pandemic management.
This project employs alternative data sources and machine learning techniques, specifically the Extreme Gradient Boosting (XGBoost) algorithm, to evaluate credit risk for individuals lacking traditional credit history. By incorporating diverse data points and addressing class imbalance through Synthetic Minority Oversampling Technique (SMOTE).
Using advanced machine learning techniques, this project successfully predicted the burned area of forest fires in the northeast region of Portugal based on meteorological and other data. By employing linear and polynomial regression models, the developed solution effectively captures the complex relationships between variables.
This project employed advanced data wrangling techniques in R to analyze and visualize greenhouse gas emissions from countries party to the UNFCCC. Utilizing the dplyr and ggplot2 libraries, the transformed data provided valuable insights into emissions trends, serving as an essential resource for decision-makers and stakeholders.
Demo
Fine-tuning GPT-2 models with custom text corpora, utilizing Hugging Face's Transformers library and advanced training techniques for sophisticated text generation applications.
hellow world
A Python implementation for image classification using a pre-trained ResNet-18 model from torchvision. The input image undergoes a series of transformations, including resizing, center cropping, tensor conversion, and normalization, before being fed into the model. The model then predicts the class label for the input
Utilizing cutting-edge machine learning techniques and advanced telematic data, a highly accurate predictive model was developed for insurance claim assessment, leading to more informed risk evaluation and optimized decision-making in the insurance industry.
Utilizing geospatial data and sophisticated machine learning algorithms, specifically K-Means Clustering, the project successfully pinpointed an optimal location for a novel Indian restaurant in New York City's dynamic landscape by evaluating competitive density.
Utilizing advanced NLP techniques and SentimentIntensityAnalyzer from the NLTK library, this script analyzes Google Play Store app reviews to extract and visualize user sentiments based on pre-defined topics, such as app interface and load time, offering valuable insights into user experience.
This project successfully integrates AWS Simple Queue Service (SQS) with a local PostgreSQL database, processing and securely storing user login data. The implementation demonstrates a streamlined approach to data ingestion, masking, and storage, resulting in an efficient and secure data pipeline.
The final code is a Python script that retrieves and analyzes financial metrics and growth data for a list of stock symbols from the Seeking Alpha API, and combines this information into a DataFrame for further analysis and visualization.
project that utilizes Generative Adversarial Networks (GANs) to generate synthetic credit data for individuals with limited or no credit history, and predicts custom credit scores based on this data using machine learning models. This project aims to improve the accuracy and inclusivity of the credit assessment process.
This repository is for my personalized GitHub profile introduction page, utilizing HTML and CSS to create a visually informative layout.