Big Data: International Relations Analysis

Project Overview

Analysed the progression of international relations between the United States and the rest of the world by performing sentiment analysis on The New York Times articles dataset. A secondary objective was to validate that sentiment against real-world stock performance data.

Articles were retrieved from the NYT Articles API, grouped by geographical location, and correlated with historical stock data from the Yahoo Finance API. Stanford CoreNLP drove the sentiment engine; Hadoop handled distributed processing; MongoDB served as the backend; and R produced the final visualisations.

Twitter Dataset Analysis & Modeling

Project Overview

Ingested, modelled, and benchmarked a large-scale Twitter dataset of 284 million relationships, 3 million user profiles, and 50 million tweets (collected May 2011).

1. Data Ingestion: Cleansed raw data and transformed it into storage-ready format.
2. Data Model Design: Evaluated reference, embedded, and hybrid document models in MongoDB.
3. Query Evaluation: Ran standard queries against each model and measured retrieval performance.

Python Malware Research

Project Overview

A security research project building a proof-of-concept Python malware that infects other Python files by appending malicious code. The malware monitors user activity, embeds scripts in affected files, makes registry entries, and disables system safeguards like Task Manager and UAC, executing itself on startup.

E-Commerce Front End: Home Appliance Reseller

Project Overview

Complete front-end redesign for a home appliance reseller. The client needed a modern website with full e-commerce functionality (product browsing, cart, and checkout), replacing their outdated static site.

Live site landing page snapshot. Click to visit.

Tutor Me: Peer Tutor Scheduling System

Project Overview

A peer tutor search and scheduling system for university students. The platform suggests tutors based on academic and personal requirements, manages profiles for both students and tutors, and handles scheduling with conflict-free room allocation.

Control Flow Diagram

ER Diagram

Web Transition Diagram

Cloud Computing Projects

Project Overview

1. Built a Combiner plugin for Hadoop WordCount, reducing map-to-reduce data transfer.
2. Extended Hadoop-Blast to run distributed BLAST bioinformatics queries across the cluster.
3. Implemented PageRank using MapReduce over a transformed adjacency matrix.
4. Loaded WordCount data directly from HBase instead of HDFS.
5. Built an HBase FreqIndexBuilder to create a full inverted index from the ClueWeb09 dataset.
6. Implemented parallel KMeans clustering using the Twister MapReduce framework.
7. Combined the inverted index (step 5) and PageRank (step 3) into a working search engine.

Xinu OS on Raspberry Pi

Project Overview

1. C++11 Future/Promise: Implemented the future-promise concurrency primitive in Xinu. Basic support  |  Advanced support
2. Custom Filesystem: Built a Unix-inspired filesystem with create, open, close, read, and write, backed by an inode table and buffer cache. Implementation details
3. Custom Unix Shell: Supported long listing, I/O redirection, pipes, and background processes.
4. Producer-Consumer: Implemented process synchronisation using semaphores.

Securing AODV in Mobile AdHoc Networks

Project Overview

Designed and integrated a game-theoretic security extension (AODV-NEW) into the reactive AODV routing protocol to defend against blackhole attacks in MANETs. Using non-cooperative game theory, AODV-NEW significantly reduces dropped packets when malicious nodes are present in the network.

MANET topology: M1 and M2 are malicious (blackhole) nodes; S is the source; D is the destination.

IU School of Education: Online Assessment Course

Project Overview

1. Built a course participant dashboard displaying student status and module completion.
2. Implemented activity-based notifications for comments and course events.
3. Configured an instructor announcement system.
4. Developed a digital badge issuance system tied to student performance.
5. Enabled course archiving so future participants can review past activity.
6. Integrated Mozilla's Open Badges API into the Open EdX platform.
7. Performed Google Analytics integration, code optimisations, and defect resolution.