I was playing around with different types of data at work when I found an interesting article on the application of time-series forecasting to predict the future performance of sports players. Of course, I jumped at the opportunity to have fun with another data science tool and apply it to Brazilian Jiu Jitsu. What is Time-Series Forecasting? Time-series data is the collection of date and time intervals, usually in a sequential order. Time-series analysis explores how something changes over a period of time and extracts meaningful insights. Time-series forecasting takes historical data, like trends by the day, month, and year, and uses a machine learning model to predict future data points based on past performances. It takes into account seasonality patterns that occur at similar intervals. In the Jiu Jitsu world, there are clothing purchases right after someone gets promoted to a new belt or a spike in instructional video sales right after a fighter wins a big tournament. Bo...
I'm a Data Scientist by day, Jiu Jitsu student by night. BJJ Black Belt at Crazy 88 in Baltimore, MD. 9 years of competing at the IBJJF Masters World Champion = 5 Gold (2020 Brown, 2018 Purple, 2017 Blue, 2015 Blue Weight +Open), 3 Silver, 2 Bronze Open. 2019 IBJJF Adult World Medalist Purple Belt. As a data scientist, I work in computer vision, natural language processing, and geek out over sport stats. This blog brings a data-driven perspective to available player, match, and tournament data.