News for hackers. Stuff that matters.
Re:Ebola death toll rises to at least 2,296: WHO by TrainedMonkey on Tuesday September 09, 2014 @05:01PM (#8293105)
There appears to be a recurrent cycle in this graph[0], is it real, or am I reading too much into unfiltered data?

[0]http://en.wikipedia.org/wiki/File:Deseased_per_day_Ebola_201...

Re:Ebola death toll rises to at least 2,296: WHO by phkahler on Tuesday September 09, 2014 @05:30PM (#8293314)
One should never differentiate a sampled signal without a low-pass filter (at least second order is preferred). The original cases/day plot was smooth, but also leveled off at the end. IDK if the people collecting the data are on any kind of weekly schedule, but that type of thing would certainly come through in the data with enough samples and Fourier analysis.