I'm disappointed but not surprised by what Steve says in his article. It's my experience too.

IMO "Data Modeling" is an extremely useful skill. It fits well with lots of other skills (like programming, database design and systems design) but is something different from all of these. Perhaps that is one reason why is seems to be overlooked and undervalued.

As @alex Gay implies, it's a really "portable" skill. The ideas of data modeling are just as valid with a small Access database as they are with a multi-zillion row, hundreds of tables ERP system (and the other way round). It's application isn't even limited to relational databases. I've used the same techniques for hierarchical structures and I've even successfully applied the principles to entirely manual filing systems. Understanding the compromises implicit n the data model can help you understand the limitations of any system derived from it. Of course, the impact and cost/benefit of those compromises is greater with the larger, more complex systems.

Data modeling doesn't require lots of fancy tools or remembering lots of complex syntax, what it does require is abstract thought and then rigorous documenting of the results (the tools are _really_ useful for maintaining the results). Maybe those are further reasons why it is unrecognised or unfashionable.

Maybe the root of the problem is that many people don't recognise "data modeling" as an activity at all. As a result they "don't know what they don't know" and so don't even think of looking for books of the kind @alex Gay mentions. This is all rather sad, especially as we are supposed to be moving to a world of "Big Data".