Creating data driven content from Python models that:

Author

Kareem Powell

Creator

    1. Predict Football Match Winners: Football underdog forecasts are generated for matches across Europe's top five leagues. Resulting content is based on weekly predictions and models' percentage accuracies when tested against prior seasons’ data. I dive into reasons behind models’ performance, comparing predictions with actual results and evaluating ways to improve accuracy—bringing you along as I test data science techniques for refining model performance, while providing you the reader an intuitive understanding of how math and statistics can be applied to real-world sports.

    2. Execute Systematic Trades: Using Warren Buffett’s value investing approach, I created a “Value Index” that comprises stocks trading at or below intrinsic value. Company stocks are evaluated on earning power, balance-sheet strength, sustainable competitive advantages, and other factors. To refine the selection, I also apply statistical methods to filter out stocks based on the relationship between company multiples. The results of this approach are measured against major U.S. indices—the market-cap weighted S&P 500, Russell 2000, and Wilshire 5000, the growth-tilted Nasdaq Composite, and the price-weighted Dow Jones Industrial Average. From there, I assess whether the approach produces higher or lower long-term returns relative to these indices, the market conditions under which it is most effective, and how to enhance my approaches using certain statistical techniques; providing readers insight into how value investing fundamentals and systematic methods can be applied to real-world investing decisions.

    • Extra Stuff: I also write quantitative consulting and personal development blog posts (see "Personal Blog"), and work on equity research and data analysis projects, from time to time (see "Projects").