OLAP cube powering Excel Pivot. What's a better solution? -
i'm looking build dynamic data environment non-technical marketers. want provide large sets of data in excel pivot table form marketers without analytics/technical backgrounds can access relevant performance information. i'm trying avoid non-excel front ends since don't want users have export data when need manipulate in way.
my first thought throw olap cube populated pre-aggregated data, got pushback team olap "obsolete." don't disagree them - there faster data processing architectures out there.
so question this: there other ways structure data marketers can access still manipulate degree in excel? i'm working 50-100m rows of data , need ability scale dimensionality.
this thoughts.
really question thrown @ team. first thought throw olap cube. didn't this. if they're achingly hip consider olap "obsolete", suggest better, more up-to-date alternative?
or, put different way - substance of objection olap solution? (i'm assuming there 1 beyond "ms gave awesome presentation of powerpivot/azure tabular, great free snacks , coffee").
your requirements pretty clear:
- easy access non-technical people
- structured data don't have interpret raw data
- access through excel
- scalability
i'll paying close attention other answers question, because i'm interested in finding out don't know something; haven't come across better solution these requirements olap.
what makes me suspicious of "post-olap" sentiment related point (2) in list above. non-technical users can tend think of cube data consume being somehow effortlessly produced, kind of magic. in indicator of success, demonstrating how easy users want well-designed olap system.
but effortlessness illusion: structure raw data form takes design effort, , resulting structure incorporates design decisions , assertions: how can easy use, because hard stuff has been encapsulated in cube design.
i have definite han solo-like bad feeling "post-olap": amounts pandering illusion of effortless transformation of data usable form, , propagates further illusions.
under olap, users wonderful magic usable data structure, , hard work done out of sight developers or me. perhaps wrong can't see data they'd - @ least users can talk , ask want.
my impression of "post-olap" sales pitch tries dispense design work. don't need pesky expensive developers, don't need make specific design decisions (which enable functionality while precluding other functionality), don't need cube-processing time-lags. can somehow deliver this:
- input data like. don't worry if it's unstructured or full of dirt!
- any scale
- immediate access analytics without etl/processing delays
- somehow, output usable, structured data. structured by... no-one in particular. user can structure like, somehow easy
call me cynical, sounds magical thinking me.
Comments
Post a Comment