Rubber Meets the Road: What’s hard about deriving real meaning from data
Details
The discipline of data science has become so popular, and celebrated lately, far outside of just the technology community — with data science “celebrities” like Nate Silver in sports and politics, applications in medicine and epidemiology, consumer marketing and more.
But in many ways this discipline is still very early in its development and has a long way to grow. So this a terrific time to have a “rubber meets the road” discussion about what is quite difficult about getting “real meaning” out of data science processes.
We'll be having 2 speakers for the evening:
- YY Lee, the Chief Operations Officer at FirstRain, will discuss how they take the perspective of deriving business developments, implicit relationships, and "meaningful insight" from vast swaths of online and social content across the global web. This means we end up having to grapple with tricky technical issues like…
- Real-world levels of variability and breadth
- Massive inconsistency and sheer messiness within real-world data sets and unstructured content
- Tough trade-offs and balancing acts: Precision vs. Recall; Emergent trends (significant, important) vs. Anomalies (just plain wrong); etc.
- “Last mile” human needs to find information relevant — elasticity, adaptiveness, contextual awareness
- Adam Baron, Director of Big Data Quantitative Research for StarMine. StarMine, a Thomson Reuters brand, first started using Hadoop in 2011 and built a home-grown quantitative finance research environment heavily using MapReduce, Hive and Mahout. In late 2014, they started using Spark (especially Spark MLlib) and have been quite impressed. Adam will discuss how StarMine uses Hadoop/Spark for researching quantitative finance models and why Spark looks like a promising new frontier.
Schedule:
Drinks and snacks start at 6pm
Talks start from 6:30pm onwards
Event Host
Thomson Reuters