Much has changed according to the possibilities how data can be stored, that will be used to answer analytical questions. Relational database management systems (RDBMS) have dominated the area of data management systems for data warehousing for decades and have been the foundation for Business Intelligence systems for long.
The row based storage approach is now accompanied by a column based approach. Disk based data storage is challenged by in-memory data storage (here in-memory does not mean, that data is not stored permanently). With the advent of Big Data solutions, it's no longer just about the decision if a RDBMS is used or one of the NoSQL databases (does the decision always mean it can just be one or the other).
Now we also have to answer questions like this: do we use Avro or Parquet (if you never heard these words before, don't worry - sooner or
later, this site will give provide some guidance when to use what).
Is MSFT Excel a valid data store that can play its part in any analytical pipeline? This is another question we have to answer, if we are architecting a pipeline (I believe it's still the most adopted tool to answer analytical questions).
Most of the time, this part of this site is about databases that are used to store data in a particular form - optimized to answer a certain kind of questions.