An essential part of building customer-facing analytics is to expose the right data to the right users, using vocabulary that they understand. This is where data modeling comes in.
Data modeling maps your databases’ tables and columns onto vocabulary that your team and your users will understand. E.g. at Trevor.io we have a table in our database called datasource_views and another one called series, both of which mean very little to our users. Once joined together and tidied up a little, however, we can comfortably refer to the end result as saved queries which is a concept that all our users are familiar with. This is data modeling.