Building a Diversified Investment Portfolio with Data Science

• In-house data science contest from Data36.com

Executive summary

• Investment experts often recommend that small investors build a well-diversified portfolio to mitigate risks. By diversifying, the impact of significant price fluctuations in individual assets or asset classes can be balanced by the performance of others.
• In this project, the task was to download data for 20 financial instruments from Yahoo Finance and perform correlation analysis. The goal was to select four assets that exhibit minimal correlation with each other, thereby creating a well-diversified portfolio.
• ⚠️ Important Note: This project aims to deepen Python programming and data analysis skills rather than to represent a viable investment strategy.

Project status

• The project is successfully completed, but there are additional opportunities to improve portfolio efficiency, for example, by applying machine learning models.

Data source

• Yahoo Finance API: https://finance.yahoo.com/

Methods used

• Data Cleaning: Preparing and organizing raw data for analysis.
• Data Table Construction: Structuring the data for further analysis.
• Statistical Analysis:
• Data normalization to ensure uniform scales.
• Correlation calculations to assess relationships between assets.
• Visualization:
• Displaying a correlation matrix for better interpretability (using seaborn).
• API Integration: Automating data retrieval from the Yahoo Finance API.

Technologies

• Programming Language: Python
• Development Environment: Jupyter Notebook
• Libraries Used:
• yfinance, pandas, itertools
• sklearn.cluster, plotly.express
• seaborn, matplotlib.pyplot, flynote

Project Results

• The analysis results and the selected portfolio are summarized in a presentation.
• Presentation: [Link to the presentation]

Origanized by

• Data36.com and dataklub.hu