This is awesome. How is the video to data conversion done? I saw OpenCV reference in the source code. Is it the tool used or something else?

I'm planning another video about the video to data conversion script. In short it's written in Python and uses various free Python libraries including OpenCV and Pytesseract.

The component was hacked together and the source code is not pretty. Please don't take it as a reference for a well designed or written program

Though it does process the videos from YouTube @vfans channel quickly and can process about 85% of the videos on that channel successfully now.

This is a nice website, I've already used the video search feature to find footage for specific matchups, especially Taka since he's not the most common character around

Thanks for very interesting analysis, SGZ.

What was the breakdown of the sample size, by the way?

Thank you for the positive feedback!

What exactly are you looking for in "breakdown of the sample size"?
You can see the data broken down in the various ways throughout the VirtuAnalytics site.

"the source code is not pretty" this is the most common phrase right after "it works for me/on my machine" so relax this is how things work, good job

Oh sorry. I must’ve missed the first 3 seconds when I watched the first time.

No problem! BTW we are at about 12.5K matches now