Digital House, Data Science
Digital House Brasil
Oct 2020 - Jun 2021 (9 months)
Optimized Brazilian Stock Portfolio Project Summary (2015 - 2020)
Objective
This project aimed to create a predictive model to identify the most profitable Brazilian stock portfolio from 2015 to 2020 using machine learning and portfolio optimization techniques. The goal was to achieve a return that would outperform the Ibovespa index, the benchmark of the Brazilian stock market.
Results Overview
Our final predictive portfolio outperformed the Ibovespa by 12% over the test period (2019-2020). Top-performing sectors included technology, healthcare, and energy, with standout stocks from companies X, Y, and Z. These selections provided consistent returns and demonstrated lower volatility compared to other options.
Exploratory Data Analysis (E