I would not bring overly complicated examples. Show us how we would add a new computed column, group, aggregate, type conversion manipulations, etc.
Anything that helps us to understand what it is we can do in shortest format possible. Current text IMO doesn't do that. On Tue, Jun 17, 2014 at 6:21 PM, Saikat Kanjilal <[email protected]> wrote: > I will refocus the blog on concrete use cases tied into the workflow of > using a Dataframe within mahout oriented perhaps towards a set of > algorithms focused on clustering, does that sound like a reasonable first > step? > > Sent from my iPhone > > > On Jun 17, 2014, at 6:15 PM, "Dmitriy Lyubimov" <[email protected]> > wrote: > > > > This blog IMO desperately needs conrete use examples, not apis. > > I would also remove R-specific examples as inconsequential. > > > > > > On Tue, Jun 17, 2014 at 6:05 PM, Saikat Kanjilal <[email protected]> > > wrote: > > > >> Hi Folks,I'm currently adding some bits to an earlier commit I had done > >> for Mahout-1490 to add a thin set of APIs around a dataframe, I was > going > >> to refine the blog specified here ( > >> > http://mlefforts.blogspot.com/2014/04/introduction-this-proposal-will.html > ) > >> to include these APIs but before I put any more effort and time into > this > >> as well as into the code I wanted to ask if there's actual interest in > the > >> developer/committer community to support this and move it into mahout. > >> Please let me know thoughts and feedback and a general direction so > that I > >> can either continue or redirect these efforts per the current need. > >> Thanks in advance. >
