Cryptocurrencies

Challenge Overview

Purpose:

The purpose of this analysis is to use unsupervised machine learning techniques is used for, preprocessing the data for PCA, reducing data dimensions using PCA,
clustering cryptocurrencies using K-means, and visualizing cryptocurrencies results.

Resources

• Software:

  • Jupyter Notebook 6.4.6
  • Machine Learning
    • Python
      • scikit-learn library

• Data source:

  • crypto_data.csv

Results

Deliverable 1: Preprocessing the Data for PCA

• Preprocessing the Data

  • Keep all the cryptocurrencies that are being traded

  • Keep all the cryptocurrencies that have a working algorithm

  • Remove unnecessary features

  • Remove null values

  • Keep the rows where coins are mined

  • Create a new DataFrame that holds only the cryptocurrencies names for using later

  • Drop the 'CoinName' column since it's not going to be used on the clustering algorithm
    

  • Use get_dummies() to create variables for text features

  • Standardize the data with StandardScaler()

Deliverable 2: Reducing Data Dimensions Using PCA

• Reducing Data Dimensions
  • Use PCA to reduce dimension to three principal components
  • Create a DataFrame with the three principal components

Deliverable 3: Clustering Cryptocurrencies Using K-means

• Clustering Cryptocurrencies

  • Create an elbow curve to find the best value for K
    

  • Initialize the K-Means model, Fit the model, Predict clusters

  • Create a new DataFrame including predicted clusters and cryptocurrencies features

Deliverable 4: Visualizing Cryptocurrencies Results

• Visualizing Cryptocurrencies

  • Create a 3D-Scatter with the PCA data and the clusters
    

  • Create a table with tradable cryptocurrencies

    • There are 532 tradable cryptocurrencies

  

  • Scale the data to create the scatter plot with tradable cryptocurrencies

  • Create a new DataFrame that has the scaled data

  • Create a 2D-Scatter