Data was gathered from the Twitter trends search on Google, Twitter trend search on Bing, Google trend search on Google, Google trend search on Bing. This data was stored and passed to VirusTotal to determine which of the URLs were malicious and which ones were not.

With the rise in popularity of cryptocurrency, crypto-mining has become an ideal form of making digital currency that can be used to buy goods and services. Crypto-mining is defined as “the process of generating crypto-currency through the use of computational power to solve complex equations as proof-of-work.” One instance of crypto-mining can be carried out with JavaScript-based programs that are embedded in a website’s code so that any visitors of the site can provide part of their processing power to mine this type of currency. This is beneficial considering that crypto mining requires an abundant amount of processing power. Alas, it has also led to crypto-jacking where a target's machine is infected with mining malware with neither consent nor knowledge. Hence, the organizations must come up with ways to ensure that they can spot the signs and shut down the miners’ illegal actions if they don't want to suffer overtaxed networks and angry customers. This paper will explore a potential solution for detecting such malicious code through binary visualization with Convolutional Neural Networks. The datasets of benign and malicious javascript codes were converted into color images using Hilbert and Entropy visualization techniques. It enabled us to train and test our CNN models to detect whether our binary image is benign or malicious. We were able to achieve 95.3% accuracy using AlexNet with Hilbert technique.