Ecosyste.ms: Awesome

An open API service indexing awesome lists of open source software.

Awesome Lists | Featured Topics | Projects

https://github.com/zara5961/cryptoclustering

This project applies unsupervised learning techniques to cluster cryptocurrencies. Using K-means clustering and Principal Component Analysis (PCA), the analysis identifies optimal cluster groups and visualizes results to uncover potential market patterns.
https://github.com/zara5961/cryptoclustering

clustering jupyter-notebook kmeans-clustering pca unsupervised-learning unsupervised-machine-learning

Last synced: 2 months ago
JSON representation

This project applies unsupervised learning techniques to cluster cryptocurrencies. Using K-means clustering and Principal Component Analysis (PCA), the analysis identifies optimal cluster groups and visualizes results to uncover potential market patterns.

Awesome Lists containing this project

README 

        

![Project](Images/Analysis.jpg)

# CryptoClustering Project



In this project, I use Python and unsupervised learning techniques to cluster cryptocurrencies and analyze if they are affected by 24-hour or 7-day price changes.



![Project](Images/unsuper.jpg)

## Project Structure

This project consists of a Jupyter notebook, `Crypto_Clustering.ipynb`, and a dataset, `crypto_market_data.csv`, containing cryptocurrency market data. I used the data to apply clustering and dimensionality reduction techniques and then visualize the results.



## Requirements Technologies

- `pandas`

- `hvplot`

- `scikit-learn`



## Instructions



### Step 1: Load and Explore the Data

1. After loading `crypto_market_data.csv` file into a DataFrame, I

2. Display summary statistics and plot the data to understand its structure.



### Step 2: Data Preparation

1. I used `StandardScaler` to normalize the data.

2. Created a DataFrame with the scaled data, setting the `coin_id` column as the index.



### Step 3: Finding the Best Value for k (Using the Scaled DataFrame)

1. I used the elbow method to find the optimal value of `k` for clustering.

2. Then Plotted inertia values for k values from 1 to 11, and

3. identified the best `k` value based on the elbow curve.



![Project](Images/image.jpg)

### Step 4: Cluster Cryptocurrencies with K-means

1. I initialized the K-means model using the best value for `k`.

2. Then Fitted and predict clusters for the scaled DataFrame and add a new column with predicted cluster labels to the DataFrame; then

4. Used `hvPlot` to create a scatter plot to visualize clusters.



![Project](Images/PCA.jpg)

### Step 5: Optimize Clusters with PCA

1. I used PCA on the scaled DataFrame to reduce features to three principal components, then

3. Created a new DataFrame with the PCA data, setting `coin_id` as the index.



### Step 6: Find the Best k for the PCA DataFrame

1. I used the elbow method to determine the best `k` using the PCA data, then

2. Plotted inertia values for k values from 1 to 11 and compared the optimal `k` value found here with the original data’s optimal `k`.



### Step 7: Cluster Cryptocurrencies Using the PCA Data

1. After, I initialized, fitted, and predicted clusters using K-means on the PCA data, and

2. Added cluster labels to the PCA DataFrame and created a scatter plot to visualize.



### Step 8: Composite Plots and Analysis

1. Created a composite plot comparing the elbow curves from the original and PCA data, then

2. Created a composite plot comparing clusters from the original and PCA data.



## Conclusion

This project explores clustering and dimensionality reduction using K-means and PCA, providing insights into cryptocurrency market behaviors based on daily and weekly price changes.