3 ms·
This is exactly something I've been looking for to experiment with user segmentation, but crucially for fast, ad-hoc segmentation. Seems to bear some similarity
by cejast 9y ago
This is exactly something I've been looking for to experiment with user segmentation, but crucially for fast, ad-hoc segmentation. Seems to bear some similarity to what Facebook [1] does for its audience insights.
[1] https://code.facebook.com/posts/382299771946304/audience-insights-query-engine-in-memory-integer-store-for-social-analytics-/ https://code.facebook.com/posts/382299771946304/audience-ins...
- jaffee 9y agoUser segmentation was actually the original use case that prompted Pilosa's development! Segmenting hundreds of millions of users with tens of millions of potential attributes.
- menegattig 9y agoVery nice job. Looks like we were trying to solve the same problem (user segmentation), in the same industry (DMP), at the same time (2013-2016). LOL. I'm the founder of a DMP too. https://blog.slicingdice.com/why-we-built-slicingdice-1beffc643571 https://blog.slicingdice.com/why-we-built-slicingdice-1beffc... I will email you guys.
- menegattig 9y agoSlicingDice founder here. We built SlicingDice exactly for this kind of necessity, very fast user segmentation. Actually, we just developed it because we didn't find any other solution that could support our needs. https://blog.slicingdice.com/why-we-built-slicingdice-1beffc643571 https://blog.slicingdice.com/why-we-built-slicingdice-1beffc...