Presentation: TweetStartups that do Smart Big Data Analysis

1. Short Talk by Knut Nesheim: "Big Data analytics: Immutability and idempotence"

We have built a streaming analytics service to aggregate a wide range of metrics. With over a billion events every day, some of our architectural choices turned out great and others turned out to be mistakes. I'll show why implementation details such as immutability and idempotence play a big role in analytics.

2. Short Talk by Mikio Braun: "Why you don’t want your realtime analytics to be exact"

Big data analytics is an expensive operation. Going for realtime even more so, as existing solutions solely rely on scaling out to achieve the necessary speed. However, many applications which can benefit from improved recency of the results do not require exact results. Examples include ad targeting, recommendation, and monitoring. This opens a potential to achieve the required speed at much smaller cost and complexity. At streamdrill we help our customers bypass these pains of scaling to realtime with a unique algorithmic solution that trades exactness for speed. This talk will explain our motivations behind streamdrill and compare it with other big data tools. 

3. Short Talk by Douwe Osinga: "Building a smart travelguide: Triposo's suggestion engine"

Triposo is a living, breathing travel guide for mobile phones that covers the entire planet. Started by two Googlers it uses clever algorithms to distill travel guides for the entire world out of content crawled online. This talk will focus on how you go from a general web crawl to extract features that let us give fresh & timely suggestions to people that are in a destination.