If i have a set of managed sources do i have to have a filter? How does that interplay work together if it does.
You do have to filter. You consume data by subscribing to a CSDL hash, via a live stream or Push subscription. essentially, Managed Sources pull the data you are interested in from Facebook/Instagram, and push that data into your platform feed, allowing the data to be collected by your CSDL filters.
The CSDL filter you define can be as simple as
source.id == "<YOUR_SOURCE_ID>" or
interaction.type == "instagram" to return everything, or something more similar to your other CSDL filters to match only content you are interested in.
Why wouldn't Youtube have a managed sources section as well then?
We have a live Youtube data feed being pushed into DataSift. Managed Sources are a method of collecting data from sources which do not offer these kinds of live feeds.
DPU : From my understanding the more operators you're utilizing within the filter section causes the DPU to go up because its a complex search. Why even offer different functions if its going to cost more?
There are a couple of factors that affect the DPU cost of a filter. These can be additional operators, use of more complex operators, or filtering for more keywords.
We provide a variety of operators to give our customers the freedom to design concise searches, to return only the data they really care about.
If you have a CSDL definition you're having trouble with, or you'd like help optimizing, please let us know via either this discussion forum, or starting a private discussion at support.datasift.com.
I've included some links to various documentation pages below which I think you may find provide some more context around my answers: