It's difficult to tell what the end goal is, but you probably need to see some of these systems and broaden your understanding first.

For example, Tokyo Metro has many out-of-gate transfers, where the user taps out of special gates (Ikebukuro M/F) and into another paid area (Ikebukuro Y) on the same fare, as long as it's within 30 minutes. There are also out-of-gate transfers between different stations, such as Shintomicho and Tsukiji. Beijing has transfers within the same paid area that are notorious for being long and/or congested, so much so that some believe it's an intentional traffic control mechanism. These all have significant implications on actual experienced commute time and fare.

So if the end goal includes human navigation or commute statistics, it's incomplete without such considerations. Those government APIs you're using are only concerned with ridership and not these details. Short of visiting these places, you could watch vlogs (there's probably 5 for every Tokyo Metro station), and you could perhaps simulate some routes on existing navigation systems.