I Turned 200+ Chinese Character Confusions Into a Structured Dataset

I'm building a Chinese learning project called HanziHero, and one feature started with what looked...