What is cspb_4502_project?

This project is a data analysis study that looks at real-time cryptocurrency prices and tries to answer three practical questions: how do different crypto assets move in relation to each other, can we automatically spot sudden price jumps or drops, and can a computer model predict whether prices are about to go up or down in the short term? The code pulls in high-frequency price data for Bitcoin and other major digital assets. It then runs a few different analyses. First, it checks how closely prices track each other. Second, it uses a statistical trick to flag moments where the price behaves unusually compared to its recent pattern. Third, it trains a type of machine learning model designed to look at sequences of data to classify whether the next short-term movement is likely bullish or bearish. The findings are straightforward. Bitcoin and Ethereum prices tend to move together, while Bitcoin and XRP do not. Over a three-day window, the anomaly detector flagged 183 unusual Bitcoin price events. The trend-prediction model got it right about 63% of the time and was better at spotting upcoming price increases than decreases. Someone who might use this includes crypto investors, traders, or anyone building tools for digital asset markets. For example, an investor could use the correlation findings to build a more balanced portfolio, while a trader might use the anomaly detection as an early warning system for risk alerts. The trend classification model is framed as a starting point for algorithmic trading, not a finished product. The README notes that the trend prediction model achieved 63% accuracy, which is modest but understandable given how noisy crypto markets are. The project is built in Python and is clearly an academic exercise rather than a production trading system, but the approach and findings are a useful foundation for someone wanting to build smarter crypto analysis tools.