CryptoClustering

Showcase the use of unsupervised learning to predict if cryptocurrencies are affected by 24-hour or 7-day price changes.

Repo Description/Outline:

Using crypto price data in a csv, data is scaled/transformed for use in unsupervised machine learning. Some other forms of data manipulation include:

• Finding the best value for k using the KMeans algorithm
• Making predictions and placing them on a scatter plot using hvplot
• Using the PCA algorithm to compare filtered data

Visualzations of these relationships can be found in the output folder.

Repository Contents:

• Resources folder containing:

  • crypto_market_data.csv

• Output folder containing:

  • plots created in the jupyter notebook

• Jupyter Notebook used to perform the machine learning

Notes & Resources:

Dataset used was provided by Rutgers University in class files