
Jeff Kelly27














So let’s kick things off. #BigData and data visualization get mentioned in the same breath a lot, but how do you view the relationship between the two?

Susan Baier
Well, one striking example for me is that I do data visualization but don't do Big Data. :-)

Jeff Kelly
From my persp, data viz is critical in "data storytelling", making sense of large volumes of messy data

Susan Baier
I think Big Data really needs data viz to be helpful, because there's just so much to look at that it's easy to get lost. But data viz definitely doesn't require Big Data -- there are lots of other kinds of data that really benefit.

Jeff Kelly
@susanbaier good point - visualizing smaller data sets can be just as compelling - the value is in the insights, not necessarily the volume or type of data

Susan Baier
I would agree about the storytelling, but insights can be hard to see even with relatively little data, sometimes.

Jeff Kelly
But as the sources of data expand and the inter-relationships of data sources grows more complex, the more important data viz becomes

Austin Dahl
I think it's the aggregation that is the key part. Without aggregation big data is too much. Visualization as a field has lots of solutions for aggregation.

Susan Baier
Unfortunately I think the two have become inextricably linked in the minds of many, so folks working with other-than-Big-Data automatically exclude data viz options as relevant for them.

Alan Morrison
@susanbaier Big Data simply implies more potential tiers of analysis, e.g., a "big data" cluster analysis of social media can lead you to the small pockets of the most interesting conversations.

Susan Baier
@AlanMorrison but I think Big Data often refers to huge data sets as well. Additional challenges there.

Alan Morrison
There are only so many kinds of data visualization, (http://www.visual-li...), and the best, most broadly grasped visualizations reflect a distillation of findings, i.e., big data made small and then visualized.

Alan Morrison
Yes, then you're finding patterns in big datasets that you wouldn't find in small datasets to begin with. But the patterns aren't displayed as big data. Rather, as distilled representations.

Alan Morrison
What @austindahl calls aggregation is a part of this.

Alan Morrison
I like the @jeffreyfkelly point on storytelling, and we've discovered that creating context with the help of full-sentence labels helps, http://www.pwc.com/e...

Alan Morrison
Visual storytelling isn't easy, but it is rewarding.