Haard Shah's Projects
In this project, Yelp restaurant reviews were analysed to gain business insights using NLP techniques. Out of 6.6 million reviews, we selected only Ontario restaurants with more than 300 reviews. We divided the 1-2-star reviews as negative reviews, 3-star reviews as average review and 4-5-star reviews as positive reviews. After removing stop words and performing TF-IDF, various topic modelling approaches were used to find out relative positive, average and negative weight of topics for different restaurants. The results obtained were visualized using Cognos BI [1]. Moreover, we also compared the running time complexity for topic modelling using Latent Dirichlet Allocation (LDA) and Non-Negative Matrix Factorization (NMF). These insights would help the business owners to improve their services.
Echo Music player android app with shake to change feature, developed in kotlin
Profile Repo
Mailhouse is a java based client-server application developed using TCP connections. It provides asynchronus communication between multiple clients. Mailhouse is a three tier architecture which includes client, server and database.
We have designed a deep learning model for predicting the spread of diabetes across Nova Scotia in 2020 so that the required arrangements for tackling the consequences can be made beforehand. The predictions have been recorded on a heat map. This heat map would help the decision-makers to arrange the required number of resources and physicians. This project was developed in 48- hours as a part of Nova Scotia Open Data Challenge.
In this era of social fashion, trends are changing every moment. People love to wear designer clothes worn by their favourite celebrity. However, as the craze for the latest fashion clothes increases, the price of such apparel increases in direct proportion. βRentHvnβ allows the users to rent an attire of their choice for a given period. The ability to select various attires to suit the occasion provides an economical option to the users. Moreover, they can also donate their old clothes to registered NGOs. Hence, this helps us to explore a market that promises a great future.
Source code for Twitter's Recommendation Algorithm
This is Training and placement cell management system project developed in C# ASP.NET. This system helps university students in automating the placement process by onlline document verification, applying to companies and viewing past year statistics