Eleni is such a nice prof and great lecturer. She teaches very clearly and has very well written slides. She will often use the blackboard to go over a proof or concept more carefully. There were also two guest lectures from researchers at Google and Facebook on Ad bidding and MapReduce, which were nice treats and quite interesting. This material is not tested.
The only thing is that this class was the "Data Science" version of the normal Analysis of Algorithms class. This meant there were a lot of IDSE students, who didn't have much CS background. This really, really slowed down the pace of the class, with all the questions that were asked. If you come from a CS background, you would be better off taking the normal Algos class.
Homework was a combination of theory and programming, however, which is nice. Any language is ok, although Python is preferred, especially for the linear programming. TAs took a long time to grade the homeworks and offered little feedback.
Exams were comprehensive, and essentially tested understanding of the main results that were showed as well as algorithm design/analysis. Averages were very low, but I think that's normal for any algos class. Final is easier as long as you understand NP-completeness and reductions.