Hemang

Community Willingness and Capability Score (CWCS)

For my Junior Design course, my team's client was iParametrics LLC. The world has been constantly plagued with natural disasters. It is imperative that national and local administration takes steps to mitigate such disasters. However, to reduce unnecessary spending, it is important to find out which communities are the best to invest into for mitigation efforts.

Hence, using machine learning, this project determines the Community Willingness and Capability Score (CWCS) of different communities and displays them in an easy to understand web application.

MapReduce Infrastructure

MapReduce is a programming model and an associated implementation for processing and generating large data sets. Users specify a map function that processes a key/value pair to generate a set of intermediate key/value pairs, and a reduce function that merges all intermediate values associated with the same intermediate key.

Store with Client and Vendors using gRPC

This project implements major chunks of a distributed service using gRPC. We build a store which receives requests from different users, querying the prices offered by the different registered vendors. On each product query, the server is supposed to request all of the vendor severs provided for their bid on the queried product. Once the store has responses from all the vendors, it collates the bid and vendor_id from the vendors and sends it back to the requesting client.

We establish asynchronous gRPC communication between the store and the user client, and the store and the vendors. Upon receiving a client request, the store assigns a thread from the thread poll to the incoming request for processing.

NCAA Basketball Ranking

In the NCAA Division 1 basketball tournament, the winner is the team that scores the most points over the season. Using a dataset, factors such as the number of games played, the number of wins, offensive/defensive efficiency, rebound rate, turnover rate, and free throw rate, will be used to create a team ranking system for each season.

We created a model that predicts the top 10 teams over multiple seasons of NCAA Basketball using Forward Selection, Linear Regression, Multi-Layer Perceptron and ReLU activation.