one of the founders here. mainly, it's a lot easier to use. we've made our syntax as basic as possible, and we handle all the COM magic on the backend. this means you can get your code working faster, and it'll automatically include a lot of behind-the-scenes optimizations to make it run better.
here's one example of an optimization: if you're using a basic Win32com script, it'll freeze Excel while it's executing. DataNitro doesn't do this, so you can turn a spreadsheet into an interactive dashboard just by using a while loop. it'd take a lot of time to get that working properly using Win32com.
It allows the storage and retrieval of structured data. To me that means it can be (and is) used as a database. Whether there are better tools is beside the point.
The reality is that Excel is the cornerstone of today's financial analysis. It is the de facto tool of the trade and can benefit from expansion of its core feature set.
As someone whose job is Financial Modelling, I wonder how distributable models / spreadsheets built using this are. That would be a barrier to adoption for my use case (and also that I'm the only person at the office who knows what Python is).
"It allows the storage and retrieval of structured data."
Can you define datatypes for Excel rows and columns? Can you define constraints on the data? If you can then I say that data is on it's way to being structured. To me, Excel is good for a quick and dirty, but using it for reliable analysis is bad practice because of its fragility. But in reality, people use the wrong tools for the job all the time and their firms and clients pay the price. Excel is the standard in financial analysis I agree, maybe that's part of the reason our financial system is so fragile.