Below are the data science projects I’ve been working on in my free time. The best thing about these projects, is that they are totally mine, hence I can give the best information.
The goal is to find films that will: inspire people, challenge their world view, make them think and make them feel. The first version was the capstone project for my data science intensive course with Springboard. I’ve since moved on and now I’m looking at building a larger database of films, as I merge two independent data sources. IMDb for most of the film information, and MovieLens for the historical user ratings. I will be looking for ways to extract more fields from the IMDb data set. This is difficult however as it is in various formats and is not fully documented, hence experimentation with the data will be needed. IMDb data is crucial as this will give more features (variables) from which to make recommendations for users. When the new data is available, I’ll be undertaking further experimentation with machine learning algorithms. After this, the next stage will be to build a light front end. (see full description). (see my code here).
Predicting the 3 best players for every match in the Australian Football League (the Brownlow results)
The goal was to predict for all AFL matches (games) in the 2016 season, which 3 players will get the 3, 2 and 1 vote(s) in the Brownlow Medal count. This would be based on the data (mainly statistics) I could obtain for that season. Australian rules football is the country’s national sport. The Australian Football League, known as the AFL, is the sport’s professional league. I correctly predicted the winner, and 8 of the top 12 players. There were 594 individual predictions made – 22 full rounds, 9 matches per round, 3 players receiving votes per match. (see full description). (see my code here).
Many lively debates about the best AFL players of all-time tend to be too subjective. The aim was to give this debate a qualitative angle by analysing the most important statistic, team goals kicked. I compared all 200+ game AFL players to see who was contributing most to this statistic for their team. (see full description).