Hello and Welcome to Delta Data!
My name is Nikola Peric, thank you for joining me. I am fascinated by big data, statistics, and machine learning. Over the course of 2013 I created a fairly accurate predictive modeling system for tennis (well several to be honest, a neural net, an Elo system, and a Glicko2 system). That being said, I am still very much learning all the time, trying to improve my intuition in data analysis.
The purpose of Delta Data is to provide a record of my learning experience that may be useful to other individuals. There will be some heavy math and statistics here and there, but, ideally, even if there are equations splurged everywhere I would like to explain the summary in easy to comprehend form, for those who may be interested. Sections with advanced math or coding will be labeled as such so you can feel free to skip them if you so wish.
Currently I am working on three things: 1) revamping my tennis modeling system. The initial system was based on a flimsy CSV dataset scraped from the ATP website. Since then I have created a far more robust database with which to work from. Throughout this blog you will see explanations for each of the models as I update them, as well as my attempt to upgrade the neural network model and create a TrueSkill model. 2) Learning the math behind a lot of the tools I use – specifically linear algebra. You may see some cool tidbits or topics here and there as a result. 3) Machine learning. I love the topic, and I’m busy delving into several books on the subject as well as playing around with scikit-learn (a Python-based library for machine learning) and Kaggle competitions. As a result it won’t be a surprise if you see some of that around here as well.
I’ll try to keep up a fairly regular posting interval to keep you all interested. Feel free to talk to me in the comments and add any of your ideas or insights to the discussion. Enjoy!
Disclaimer: Writings in this blog are not the be all, end all of the subject discussed. There may indeed be a more efficient or effective way to do the things discussed that you know and I do not. Therefore I strongly encourage you to comment on the entry and discuss it – make it a learning experience for both myself and other readers.