Is there any point in NoSQL data modeling tools since many NoSQL solutions are schema-less?

JSON maps to UML quite well, you can use strong aggregation to represent embedding for example. However, its going to come down to preference of how you want to represent the schema. You make a valid point that some application are schema-less, but more typically an application is going to have a stable schema for a given release... this is when a diagramatic representation may be more useful.

Validation, type-checking, and sanitization. One use case, from MongoDB, is when the database doesn't enforce types for any given key, you have to do it the application level.  So if one part of your application was saving a user-input of a number as a string, and another part was querying it as an integer, the query would turn up empty because the query is type sensitive, but the individual documents aren't.   In practice, this could engender all kinds of bugs or bad user experiences because it's a silent failure - the database never throws an error for having input of the wrong type, and the queries happily return "incomplete" data sets that only match the value and type you're querying.  Having an object model with a defined type for each key prevents this kind of problem.  All of the other benefits of validation and data sanitization would come along with these data modeling tools.  My experience is primarily in Ruby with mongo_mapper.

Data model is a general concept and is not related to SQL or NOSQL. Data model is critical as it defines what and how data is stored and having right tool in place that enables its rapid development, modification, change management and sharing is critical as well.

I'd invite you to check out the http://databasel.com/?ref=Is-there-any-point-in-NoSQL-data-modeling-tools-since-many-NoSQL-solutions-are-schema-less&details=answer . It was designed as Business Intelligence for NoSQL.We would appreciate to knowing your business case, feel free to reach me via email :