Skip to content

Latest commit

 

History

History
88 lines (65 loc) · 3.88 KB

README.md

File metadata and controls

88 lines (65 loc) · 3.88 KB
Raw

Cryptocurrencies Analysis

drawing

Executive Summary

The main objective of this project was to determine which cryptocurrencies were trading on the market and to use unsupervised machine learning to see if there are any clusters of cryptocurrencies.

Four clusters of cryptocurrencies were used. In the dashboard below, the final page facilitates potential investors who are interesting in choosing the right mix of cryptocurrencies based on Class, Proof Type and Algorithm for their investment portfolio.

Watch the video Click above image to view video of Dashboard.

For powerBI users, view dashboard

Project Outline

  1. Perform dimension reduction using Principal Component Analysis (PCA) method and clustering using K-means
  2. Present my findings

Resources

  1. Python
  2. Sklearn
  3. Pandas
  4. HVplot

Data Preprocessing Tasks

  1. Remove all cryptocurrencies that were not trading
  2. Remove all cryptocurrencies that had no defined Algorithm
  3. Remove the IsTrading column
  4. Remove all cryptocurrencies with at least one null value
  5. Remove all cryptocurrencies without coins mined
  6. Create a separate dataframe with all the names of the cryptocurrencies
  7. Remove the CoinName column in the original crypto_df dataframe
  8. Create dummy variables for the Algorithm and ProofType columns
  9. Standardize the dataframe

Reducing Data Dimensions Using Principal Component Analysis (PCA)

Reduced to Three components

Elbow Curve Analysis

Experimented with values for K=4, K=5 and K=6

Decided that K=4 was the best option

Visualizing Results

3D plot with parameters hover_name="CoinName" and hover_data=["Algorithm"]

References

  1. dPoC
  2. PoS
  3. PoW
  4. PoS Guide
  5. PoA
  6. Other Explantions