Sounds interesting!
I can give you a very brief description, but our application may seem a bit unusual and is catered for a researcher community.
Our use case is a database for media researchers. We are creating an inventory of news sources and keep track on their ownership relations (which is the reason we use a graph database). There are also more kinds of entries (e.g., software packages for text analysis) and we are planning to grow the data model to also cover other kinds of information sources (e.g., social media accounts of political actors).
You can view it here: https://meteor.opted.eu/
I am sometimes required to report some usage statistics and that also includes the number of new entries made in a certain time period. Because dgraph does not offer a native (and convenient) function for datetime aggregations, I run this kind of analysis with pandas (using just a database dump).