NoSQL Workbench from AWS works really well. However, I found the process of setuping up tables unnecessary hard.
I came with the idea, that the table and schema would be generated from the text input. You can set up your primary key and other GSI indexes. Also, you can click on these indexes and aggregate data into partitions.
#pk, $sk, ##pk1, $$sk2, info
User-A, INFO, score, 3, Additional information
User-A, Connection
User-B, INFO, score, 4, Additional information
User-C, INFO, score, 5, Additional information
User-D, INFO, score, 1, Additional information
Table is created with your data and indexes. Take a look at header of the table. The primary key and GSI indexes are clickable. By clicking on them, you activate aggregating data by this index (also sort by the appropriate sort key). With this way, you can explore access patterns much easier.
Pretty cool hah?