CSE498, Collaborative Design, Fall 2014
Computer Science and Engineering
Michigan State University

Amazon Marketplace provides a platform for individuals and businesses to sell products to hundreds of millions of online customers. Since more than 40% of Amazon unit sales now result from third-party sellers, Amazon is committed to improving and optimizing their experiences.

One of the main challenges faced by Amazon sellers is that of determining what products to stock and in what quantities. To assist them, the Amazon Seller Services team would like to provide its sellers with a free and fast alternative to traditional market studies, which are expensive and time-consuming.

Our Twitter Trending Effects on Amazon Sellers system analyzes Twitter tweets to determine what Amazon products consumers are talking about, what they think of them, and what products Amazon customers are buying.

These Twitter trends are displayed in a responsive, real-time dashboard that Amazon sellers customize for their specific needs, adjusting timeframes and overlaying data sets for comparison. Information about a particular brand or product includes the volume of tweets, the overall positive or negative sentiment and the sentiment of tweeters who mention owning it or purchasing it.

Using our dashboard to forecast the popularity of brands and products, Amazon sellers can make better-informed decisions for the types and quantities of products they stock.

Our system is written in Java using the Kinesis, Redshift, and Elastic Compute Cloud (EC2) Amazon Web Services with a cloud-hosted Ruby on Rails web interface.