Two semesters of Analytics coursework in the Department of Information & Decision Sciences including data warehouse dimensional modeling, data mining and marketing analytics.
- Evaluation of Inter-Industry Merger Data to Predict Company Performance
- The idea behind the project was to understand the effect that strategic partnerships have on performance, especially the role that software companies have in these alliances. Started with three large sets of data of diverse formats acquired from various sources like Twitter, Wikipedia, NAICS databases and other financial databases. Cleaned data and constructed database on MySQL Workbench for report generation. 1.5M rows of data. Generated reports varying complexity for extraction of specific data points. Analyzed reports using SPSS Modeler and crafted graphs to visualize results.
- Partnerships with Software firms posted better and improved results post merger
- Correlation Between Twitter Chatter & the Financial Markets
Objective was to understand if Twitter influences trade decisions before earnings releases.Analysis of over 1.2 million rows of feed data from Twitter and Yahoo finance in order to analyze co-relation between financial earnings releases, trade volume and twitter activity. MySQL for database, KNIME for statistical analysis and Excel for visualization. Significant spikes in Twitter chatter and trade volumes before earnings releases. Text analyses of scrapped chatter was proposed to understand correlation of Twitter sentiment with Trade volume.
- Analysis of YouTube Video Metrics for Engagement Benchmarks in an Industry
The objective was to benchmark the metrics like Views, Likes, Dislikes, Video Duration and posting frequency for the YouTube channels of companies that sell Paleo Diet products. This was done for a consulting project. Scrapped video metrics like views, likes, dislikes, video duration and posting frequency for all videos of ten YouTube channels, of companies big and small that sell Paleo goods using R and Python. Analyzed data and visualized results using Tableau, Excel and amCharts.
- DSLR Photography Assistance using IBM Watson Cognitive Analytics (Android Application)
Acquired many manuals and guidance materials on DSLR photography from diverse sources. Curated content to match Watson’s input requirements. Constructed corpus to train Watson. Trained Watson on 1,000 Q&A pairs clustered around topics in the corpus to achieve greater than 88% confidence. Created Business Model, five year plan and strategies for a sustainable start-up. Assisted in design of Android application that will serve as front-end and product.
- Routing for Mobile Ad-Hoc Networks using Ant Colony Optimization Metaheuristic
Proposed a simple routing algorithm for mobile ad-hoc networks based on the ant colony optimization metaheuristic framework. Conducted literature survey, framed the algorithm and demonstrated a preliminary version using the simulation method. The algorithm was designed to emulate the foraging behavior of ants while looking for food. The ACO is a probabilistic technique for solving computational problems, which can be reduced to finding good paths through graphs.