Ask HN: Data Schema Design Process

I am designing a data-intensive web application. Because I am not exactly a full stack engineer I catch myself often re-modeling data schema I decided on as the work progresses. This effectively costs me a lot of time and back-and-forth in the project. I have a feeling it can be a more organized process than just rules of thumb all the way.

I am dealing with mixed data types. On one hand I need key-value structure, on the other hand, I need it to hold references to the filesystem/files directories. All of data comes well-structured but it can often be of considerable size.

Question is, what are good resources, case studies or personal blog posts in the area of building a data-intensive app. I am looking specifically for the resources on deciding how to organize data for storage and future retrieval/transformation in context of smaller/medium size architectures (ie. iCloud case study is of little use to me most probably)