CHLA Appointment Prediction Tool

Github LinkApp Link

This project covers the deployment of a machine learning model aimed at predicting patient no-shows at Children's Hospital Los Angeles (CHLA). The objective is to accurately predict whether a patient will show up for their scheduled appointment using historical CHLA data.

The model is developed and deployed using various methods:

1.  Model development and deployment using Azure ML Studio.
2. Developed using python (sci-kit learn); created and deployed as Streamlit app.
3. Deployed using Docker (containerization).
4. Deployed using FastAPI (and also Docker for containerization).
5. Deployed using AWS.

A variety of machine learning algorithms and hyperparameters were explored to select the best-performing model. The Random Forest Classifier was chosen due to its robustness and performance across several metrics crucial to the no-show prediction problem. Emphasis was placed on recall, precision, F1-score, and ROC-AUC, given the high costs associated with false predictions.

Activision SQL Project

Github Link

This project demonstrates advanced data analytics skills in the gaming domain, inspired by a job posting from Activision. It uses data from Kaggle and the Steam API to analyze trends in gaming, especially focusing on the Call of Duty series. This analysis is meant to showcase skills relevant for a potential analytics role at Activision.

Data Collection: Gathered data from Kaggle and the Steam API.

‍Data Cleaning and Preprocessing: Prepared the data for analysis by cleaning and structuring it using Python.

‍SQL Analysis: Primary analysis was conducted using SQL to investigate and manipulate the data efficiently.

Spam Email Detector using Naive Bayes

Github Link

This project implements a Spam Email Classifier using the Naive Bayes supervised machine learning algorithm. The classifier is trained and tested on a dataset of emails, which are labeled as either 'Spam' or 'Not Spam'.

Movie Recommendation Engine using kNN

Github Link


This project implements a simple movie recommendation engine using the k-Nearest Neighbors (kNN) algorithm. It's designed to recommend movies similar to a given movie, based on genres and IMDB ratings.

This prototype serves as the backbone for a hypothetical movie recommendation website, similar to those used by Netflix or Hulu, but on a smaller scale.

Location Synergy Project

Project Coming Soon...

This project focuses on creating a web application using React.js to showcase a research initiative led by one of my professors. The application, centered on Location Synergy analysis, is designed to optimize site selection for new business ventures.

Users of the app can input an address and select a specific industry code. The app then calculates and displays a synergy score along with other relevant metrics to aid in making informed decisions about business locations.

Capstone Project: Hawthorne Advertising

Project Coming Soon...

As part of my Master's Program at LMU, I will be working on my capstone project with Hawthorne Advertising. This project aims to enhance advertising efficacy through a media mix modeling approach using Bayesian probability.

The objective is to analyze time series data from various advertising channels and develop a predictive model that quantifies the impact of each channel on sales, including adjustments for seasonality and other variables.

Key deliverables include insights on optimizing advertising spend, a baseline sales estimation without advertising, and the development of compelling Tableau visuals to depict the findings effectively.

Income Level Prediction using ML

Decision Tree LinkEnsemble Models Link

This project involves the development, evaluation, and comparison of various ensemble machine learning models to predict income levels based on U.S. Census Bureau data.
‍
The model aims to classify individuals into two income categories: ">50K" and "<=50K", using seven demographic variables.

This comprehensive guide covers data preparation, model building, performance evaluation, hyperparameter tuning, and analysis, providing a thorough understanding of how these models can be optimized and visualized to achieve accurate income predictions.

Cloud Resume Project

Github LinkWebsite Link

This project highlights my proficiency in cloud computing, showcasing the various AWS tools I utilized to deploy a static website (my resume).

The following AWS tools were used:
- EC2 Instance
- S3 Buckets
- Route 53 & Certificate Manager
- CloudFront
- Lambda Function
- Dynamo DB & RDS
- API Gateway

To learn more about how I made the Cloud Resume, check out my article on Medium: The Cloud Resume Project.

Yelp Data Collection

Github Link

This project contains two Jupyter Notebooks designed for collecting Yelp data using two different methods: Yelp API and Web Scraping.

Whether you are interested in analyzing Yelp business reviews, understanding customer sentiments, or exploring dining options in a specific area, these notebooks provide a detailed step-by-step guide to gathering the data you need.

Nike Customer Analysis

Github Link

This project, undertaken in one of my undergraduate classes, focused on supporting Nike's sales growth through data-driven insights (hypothetical scenario). Collaborating with a team, our goal was to address Nike’s challenges with declining sales through strategic analysis. We did the project through a structured, sprint-based approach managed on Trello.

Sprint 0: Setup
‍Defined the business problem.
Conducted data collection.
Established project structure and collaboration channels.

‍Sprint 1: Foundation Building
‍Developed the ETL pipeline and data warehouse.
Implemented Data Governance policies.
Created a descriptive analytics dashboard.

‍Sprint 2: Optimization and Analysis
‍Automated the ETL process.
Performed predictive modeling and A/B testing.
Authored a data manifesto.
Developed a diagnostic analytics dashboard.

Following these sprints, we presented our findings and provided actionable recommendations for Nike to enhance sales strategies.

Web App for Retail Investors (Capstone)

Github Link

For my undergraduate capstone project, my teammate and I developed an interactive web application designed to display stock information and provide "buy" or "sell" predictions using machine learning.

The core of our predictive functionality was built around a Random Forest model.

My specific contribution to the project was the development of the web app interface using React Native. I integrated various APIs to fetch real-time stock prices, news updates, and other relevant company metrics. The application presents these data points alongside the machine learning model's predictions, offering users a comprehensive tool for informed stock trading decisions.

The project is no longer hosted on AWS due to the project scope and cost.

Data Prep for Machine Learning

Github Link

This repository contains a Jupyter Notebook for the Exploratory Data Analysis (EDA) of a comprehensive dataset on house pricing. The notebook includes various data cleaning and preparation techniques to make the dataset "analytics ready" for machine learning.

The House Price Dataset includes features potentially predictive of house prices. This data serves as a foundation for developing machine learning models aimed at predicting house listing prices, thereby aiding sellers in setting competitive prices for their properties.

Vish's Portfolio

Welcome to my portfolio website!
‍
Here you will find a curated list of my projects that showcase my skills in data analysis and web development.

Explore
About
Projects
Links

About me

Hello! I’m Vishwamshankar (Vish) Ramesh, a passionate Business Analytics master's student dedicated to leveraging data to solve complex problems and drive strategic decisions. With a robust background in data analysis, machine learning, and web development, I thrive on turning insights into actions that positively impact our world.

I am eager to connect with like-minded professionals and explore potential collaborations. Feel free to reach out through LinkedIn/GitHub/Email, or explore my projects here on my website.

CHLA Appointment Prediction Tool

A project that showcases different methods to deploy an ML model.

Activision SQL Project

Data Collection and SQL Analysis on video game data. Inspired by a job posting by Activision.

Spam Email Detector

An ML model to detect spam email using the Naïve Bayes classification algorithm.

Movie Recommendation Engine

A prototype of a movie recommendation engine that uses the k-Nearest Neighbor algorithm.

Location Synergy Web App

A web app to showcase a research project on calculating location synergy for businesses.

Capstone: Hawthorne Advertising

Master's Capstone Project on Media Mix Modeling with Hawthorne Advertising.

Income Prediction (ML)

Using decision tree and ensemble methods to predict income level.

Cloud Resume

A static website hosted and deployed using several AWS cloud computing services.

Yelp Data Collection

Yelp Review Data Collection using API and Web-Scraping.

Nike Customer Analysis

An end-to-end data analytics project based on a job posting by Nike.

Web App for Retail Investors

A React Native web app to demonstrate my capstone project.

Data Prep for ML

Data cleaning and preparation techniques to get analytics ready data for Machine Learning.

Important Links

Here are a few links where you can connect with me or explore more of my work.

GithubLinkedInCloud ResumeTableau Public