Feast on the Desperate

The Challenge: This paper addresses the escalating threat of cryptocurrency scams by examining the social engineering and psychological tactics that exploit individuals. We identify a critical gap in public awareness and expose how social media, leveraging platforms, influencers, and targeted advertising, is a primary vector for these sophisticated schemes. The research highlights that these scams succeed not just through technical means, but by preying on human vulnerabilities like FoMo (Fear of Missing Out) and other emotional manipulations.

Methodology & Analysis: Through a detailed analysis of scam victimisation and real-world testimonies, this work delves into the psychological underpinnings of why individuals fall prey to these schemes. The paper synthesises elements from disparate fields, phishing prevention, the SCAMS Checklist, the I-PACE model, and insights from parasocial relationships to develop a new, comprehensive framework for scam prevention. This interdisciplinary approach provides a more holistic defence strategy.

The Key Findings & Contributions:The study's findings confirm that scammers systematically exploit financial fears and parasocial trust to bypass logical decision making. Our primary contribution is the proposed comprehensive framework, which empowers individuals with a combined set of tools to identify and avoid fraudulent cryptocurrency schemes. This framework's novelty lies in its integration of psychological and social factors with established technical prevention methods.

Implications & Conclusion: The implications of this research are significant for both individuals and organisations. By understanding the psychological tactics employed by scammers, we can better equip ourselves to resist these manipulative strategies. The proposed framework serves as a valuable resource for developing targeted educational initiatives and preventive measures in the fight against cryptocurrency scams.

Code Red for Healthcare: Why the NHS is Losing the War on Ransomware

The National Health Service (NHS) is in a state of perpetual crisis, grappling with understaffing, overcrowding, and systemic delays. This operational fragility is being dangerously amplified by a relentless wave of cyberattacks. Ransomware, in particular, has proven to be a uniquely devastating threat, capable of crippling hospital operations, compromising patient data, and putting lives at risk.

This analysis argues that the NHS's vulnerability is not just a matter of insufficient funding but a fundamental flaw in its technological foundation. Its deep-seated reliance on traditional, high-maintenance operating systems like Windows creates an attack surface that is simply too vast to defend. The solution lies in a strategic pivot to a modern, secure-by-design platform. By adopting a 'Zero Trust' architecture, exemplified by ChromeOS, the NHS can build a more resilient, manageable, and inherently secure infrastructure fit for the challenges of the 21st century.

JPMorgan-Transaction - Detecting Financial Fraud with Machine Learning

This project tackles the critical challenge of fraud in the burgeoning world of mobile money. By analysing a large-scale dataset from a financial services provider, I uncovered the subtle behavioural patterns that distinguish legitimate customers from fraudulent actors.

The core of this work involved moving beyond existing detection flags to engineer a highly accurate predictive model. The result is a set of actionable insights and a robust machine-learning framework designed to help organisations pre-emptively identify and stop fraud, enhancing system security and protecting customer assets.

Spam Classifier using Logistic Regression

The spam classifier project that uses a Logistic Regression model to identify spam emails. Built with Python, Scikit-learn, and Pandas, the project leverages the Enron email dataset for training. The process involves cleaning the email text, converting it into a numerical format using the CountVectorizer technique, and then training the model to recognise patterns. The repository includes a Jupyter Notebook that details the training and evaluation process, which measures performance through metrics like accuracy and a confusion matrix. The author also suggests potential future improvements, such as experimenting with different algorithms like SVM or incorporating more advanced text preprocessing methods.