Standalone E-Commerce Web Application
E-commerce web application was designed and developed by exploring the Hadoop ecosystem using Node JS and NoSQL database Mongo DB was used to store the data sets. Search functionalities were implemented using Elastic Search indexing over database. Redis Cache was used to cache the search results and to store the frequently accessed data. Image hosting was done through CDN.
Tools: Hadoop, Node JS, MongoDB, CDN, Elastic Search
US Death Records
Extracted, cleaned the data and analyzed pattern of deaths in terms of age groups, diseases, manner of death, death rate per year and identified significant variables. Regression, clustering and classification models were used to predict the life expectancy and death likelihood rate. Also, developed an interactive dashboard to study the various trends in the pattern of death.
Tools: R, Shiny
Predicting US Elections 2016
Analyzed and cleaned the primary election data records and built regression models and decision tress using machine learning algorithms neural networks, naïve bayes and decision trees. Also, performed sentimental analysis by extracting data from twitter and visualizations were plotted.
Tools: Python, SAS, Rapid Miner, Tableau