aksharas28
I am a final semester Computer Science graduate student at SUNY Buffalo. Actively looking for full-time opportunities in the field of Software Engineering
Student at SUNY BuffaloBuffalo, NY
Pinned Repositories
Breast-Cancer-Classification-using-Logistic-Regression
Wisconsin Diagnostic Breast Cancer dataset was collected from the UCI machine learning repository and a machine learning model was built that can classify the cancer cells to Benign and Malignant. The data was trained using Logistic Regression and the model’s performance was tested on the test data.
Cloud-Data-Warehouses-with-AWS
In this project, an ETL pipeline is built for a database hosted on Redshift. In this project, data from S3 is loaded to staging tables on Redshift and execute SQL statements that create the analytics tables from these staging tables.
Data-Lake-with-Apache-Spark
In this project, we try to help one music streaming startup, Sparkify, to move their data warehouse to a data lake. Specifically, I built an ETL pipeline to extract their data from S3 and processes them using Spark, and loads the data into a new S3 as a set of dimensional tables. This will allow their analytics team to continue finding insights into what songs their users are listening to.
Data-Modeling-with-Apache-Cassandra
In this project, data modeling with Apache Cassandra is performed and an ETL pipeline using Python will be completed.
Data-Modeling-with-Postgres
In this project, data modeling with Postgres is complete and ETL pipeline using Python is built.
JPMorgan-Virtual-Software-Engineering-Internship
Interface with stock price data feed and Data Analysis. Using JPMorgan Chase frameworks and Perspective tool for implementing for Data Visualization. Using Perspective to create the chart for the trader's dashboard
Neural-network
Implemented a single layer hidden layer neural network in Python from scratch. Secondly, a multi-layer Neural network is built with Keras Library. Finally, a Convolutional Neural Network (CNN) is implemented with Keras Library. In the end, a comparison of Loss, Accuracy, Confusion Matrix for each classifier is made and relative strengths and weakness is observed.
Reinforcement-Learning-Q-Learning
A reinforcement learning agent is built to navigate the classic 4x4 grid-world environment. In this project, the agent has learnt an optimal policy through Q-Learning which will allow it to take actions to reach a goal while avoiding obstacles. Key components of the Q-Learning algorithm are implemented and explained. Specifically, tabular Q-Learning approach is used which utilizes a table of Q-Values as the agent’s policy.
UB-CSE535-Evaluation-of-IR-Models
The goal of this project is to implement various IR models, evaluate the IR system, and improve the search result based on your understanding of the models, the implementation and the evaluation.
UB-CSE535-Information-Retrieval-Search-Analytics
Analyzing the impact of political rhetoric in social media by creating an end-to-end website by hosting on AWS and by collecting tweets from twitter of 15 POI'S
aksharas28's Repositories
aksharas28/Cloud-Data-Warehouses-with-AWS
In this project, an ETL pipeline is built for a database hosted on Redshift. In this project, data from S3 is loaded to staging tables on Redshift and execute SQL statements that create the analytics tables from these staging tables.
aksharas28/Data-Lake-with-Apache-Spark
In this project, we try to help one music streaming startup, Sparkify, to move their data warehouse to a data lake. Specifically, I built an ETL pipeline to extract their data from S3 and processes them using Spark, and loads the data into a new S3 as a set of dimensional tables. This will allow their analytics team to continue finding insights into what songs their users are listening to.
aksharas28/Data-Modeling-with-Apache-Cassandra
In this project, data modeling with Apache Cassandra is performed and an ETL pipeline using Python will be completed.
aksharas28/Data-Modeling-with-Postgres
In this project, data modeling with Postgres is complete and ETL pipeline using Python is built.
aksharas28/UB-CSE535-Evaluation-of-IR-Models
The goal of this project is to implement various IR models, evaluate the IR system, and improve the search result based on your understanding of the models, the implementation and the evaluation.
aksharas28/UB-CSE535-Information-Retrieval-Search-Analytics
Analyzing the impact of political rhetoric in social media by creating an end-to-end website by hosting on AWS and by collecting tweets from twitter of 15 POI'S
aksharas28/Breast-Cancer-Classification-using-Logistic-Regression
Wisconsin Diagnostic Breast Cancer dataset was collected from the UCI machine learning repository and a machine learning model was built that can classify the cancer cells to Benign and Malignant. The data was trained using Logistic Regression and the model’s performance was tested on the test data.
aksharas28/JPMorgan-Virtual-Software-Engineering-Internship
Interface with stock price data feed and Data Analysis. Using JPMorgan Chase frameworks and Perspective tool for implementing for Data Visualization. Using Perspective to create the chart for the trader's dashboard
aksharas28/Neural-network
Implemented a single layer hidden layer neural network in Python from scratch. Secondly, a multi-layer Neural network is built with Keras Library. Finally, a Convolutional Neural Network (CNN) is implemented with Keras Library. In the end, a comparison of Loss, Accuracy, Confusion Matrix for each classifier is made and relative strengths and weakness is observed.
aksharas28/Reinforcement-Learning-Q-Learning
A reinforcement learning agent is built to navigate the classic 4x4 grid-world environment. In this project, the agent has learnt an optimal policy through Q-Learning which will allow it to take actions to reach a goal while avoiding obstacles. Key components of the Q-Learning algorithm are implemented and explained. Specifically, tabular Q-Learning approach is used which utilizes a table of Q-Values as the agent’s policy.
aksharas28/UB-CSE535-Boolean-Query-and-Inverted-Index
Created an inverted index, postings list, and implemented Document-at-a-time OR and Document-at-a-time AND for answering Boolean queries. The result needs to be ranked based on TF-IDF
aksharas28/Unsupervised-Learning--Clustering-Analysis-on-Fashion-MNIST-Data
To perform cluster analysis on Fashion MNIST dataset using unsupervised learning, K-Means clustering, and Gaussian Mixture Model clustering is used. The main task is to cluster images and identify it as one of many clusters and to perform cluster analysis on fashion MNIST dataset using unsupervised learning. The model’s effectiveness is measured by testing the machine learning scheme on the testing set and the performance can be evaluated by its clustering accuracy. Three tasks performed are K-Means algorithm to cluster original data space of Fashion – MNIST dataset using Sklearns library, an Auto-Encoder based K-Means clustering model is built to cluster the condensed representation of the unlabeled fashion MNIST dataset using Keras and Sklearns library, an Auto-Encoder based Gaussian Mixture Model clustering model is built to cluster