Crypto Clustering - sklearn K-Means

This challenge utilizes Python and unsupervised learning to predict if cryptocurrencies are affected by 24-hour or 7-day price changes.

‍

GitHub Repository

‍

Results

Elbow Curve

Original Data

Transformed Data

Clustering

Original Data

Transformed Data

Analysis

1. What is the best value for k?

• 4

2. What is the total explained variance of the three principal components?

• 89%

3. What is the best value for k when using the PCA data? Does it differ from the best k value found using the original data?

• 4
• PCA data, visually, follows a very similar elbow curve to the original data, best value for K-Means remains consistent between the transformed and original data.

4. After visually analyzing the cluster analysis results, what is the impact of using fewer features to cluster the data using K-Means?

• The main impact visually, seems to be higher quality clustering (transformed clusters are visually more compact and well-separated, compared to the original data scatter plot). Outliers are also more obvious with the transformed data.
• Ultimately leads to more informed and accurate interpretations of the results.

Resources

• Introduction to K-Means Clustering Algorithm
• sklearn KMeans Documentation

‍

‍

GitHub Repository

‍