Projects
Portfolio Benchmarks
The Portfolio Benchmarks project is aimed at analyzing and benchmarking the performance of a financial portfolio against major market indices like the S&P 500 and TSX300. It involves fetching historical market data, calculating performance metrics, visualizing data, and predicting future portfolio values. The project structure includes Python scripts for data manipulation and analysis, Docker configurations for containerization, and a Jupyter Notebook for interactive data visualization.
Key Features:
-
Data fetching from Yahoo Finance using yfinance.
-
Performance comparison with S&P 500 and TSX300 indices.
-
Calculation of financial metrics such as Sharpe Ratio, Alpha, and Beta.
-
Interactive visualizations with plotly.
-
Predictive analysis for future portfolio performance using scikit-learn.
-
Containerized Jupyter environment for analysis.
R Housing Price Predictions
The R Housing Price Predictions project involves the analysis of the KC House Sales dataset to predict house sales in King County, Washington State, USA, using Multiple Linear Regression (MLR). The dataset, sourced from Kaggle datasets under the name "KC_HouseSales_Data", includes historical data of houses sold between May 2014 to May 2015.
Data Source: The dataset is available on Kaggle.
​
Key Features:
-
Data preprocessing and exploration to understand dataset characteristics.
-
Exploratory Data Analysis (EDA) to identify patterns, outliers, and relationships between variables.
-
Data visualization to support EDA findings.
-
Building a Multiple Linear Regression model to predict house prices.
-
Evaluating model performance and comparing it with a one-layer forward neural network as a reference.
Libraries and Tools Used:
-
tidyverse for data manipulation and visualization.
-
corrplot for visualizing correlations between variables.
-
lubridate for date-time manipulation.
-
Various tools for model building and evaluation, including caTools, GGally, caret, and leaps.
