Every so often I will think of, or stumble upon, a small data-driven project that I find interesting to tackle given my background in research and data-analysis. Those that I deem interesting enough I will write up into small pages. On this page, I will provide a brief summary of these toy data science projects along with links to the full analysis.
The AFL Brownlow medal is awarded to the player at the end of the
season with the most votes. These votes are awarded on a 3-2-1 basis
by the officiating match day umpires. However, the outcome of these
votes are a closely guarded secret until after the home and away season,
when they are revealed at the end of season Brownlow medal ceremony.
This project aims to build a machine learning model that predicts the
3-2-1 votes based on the readily available player statistics on match day.
It is trained on over 15 years of historical data. For more details,
please head on over to the summary page of this project.
As a huge fan of LEGO, I was curious as to whether the cost of LEGO has
gotten more expensive in recent times as is often perceived to be the case.
Further, I was also interested as to whether licensed LEGO themes such as
Star Wars are more expensive than unlicensed themes. Finally, I also
decided to look into whether the cost of LEGO in Australia was cheaper or
more expensive than other major markets such as the US.
With these questions in mind I performed some in-depth data-analysis. For this exploration I built a dataset based off 35 years of historical LEGO set information (from between 1990 and 2024). If you are interested, take a look at the project page on which I detail my exploration.