To expand on what @hansen provided, snapshots and backups are full-state backups, meaning they backup your Droplet as-is, in it’s current state.
When you restore from a snapshot or backup, it restores the Droplet entirely, not just specific data. If you need to be able to recover specific data, such as a single database, one or more files, etc, then it would be best to setup some sort of backups on the server and store them offsite to block storage or another medium.
For example, if you had Apache, MySQL, and PHP on the Droplet, along with multiple databases and a few hundred files, and you restore from a snapshot, all of that is recovered. You can’t (as of this reply) pick and choose what you want to restore.
You can, however, restore the image to another Droplet, access it, and pull down the data you need. I have done that in a few cases for clients where that was all they had and restoring from a full backup wasn’t ideal or feasible. Once done, you can then destroy the Droplet that you brought online to pull data from.
When it comes to backups, the snapshot and backup service is nice to have, though I wouldn’t rely on it exclusively. You should still implement other forms of backup – whether it’s storing to block storage or sending them to another data center.