I personally think the interesting part about de-making is to try if a style and feeling can be brought down to a few pixels.
The Dark Souls style was pretty interesting because it's working with real proportions and lots of small details which aren't easy to translate to pixels. and it also needed a lot of version to get something acceptable. I am not sure though if anybody recognizes the sprite now as "Dark Souls" if it's without context. I fear that it also could be just read as a generic dark knight.
I think De-Making works best with established main-characters
I made an experiment quite a while ago where the subject was "Metroid Prime". I started with the Prequel "Super Metroid" which was my reference and I used the designs and the mood of Metroid Prime. The bad thing is that I weren't able to do more than a single sprite so far.
design ref:
http://images2.wikia.nocookie.net/__cb20100215035740/metroid/images/5/58/Mprime_07_big.jpgThe hard part of de-making something which hasn't a 2D preqel is that the gameplay of the 2D version also has to be envisioned before starting. The part of analizing the earlier 2D graphics and applying principles to new stuff won't happen - but I think from the educational side this is also pretty interesting. And it's cool to work with concept art which is intended fo rsomething completely different too.
LoL might be popular, but I personally never played it- from what I know it lacks a strong main character (there are dozens of heros), it has no 2D prequels and the environment is also very generic. I don't know about their concept art, they have digital paintings for each character, but I haven't seen turnaround drawings or environment concept art so far.
For a group activity it might be more successful to choose one single game and we try as a group collaboration to come up with at least one mockup - this wouldn't be that much work for everybody and we could analyze every aspect in a group too.
Just my 2 cents