Transcript: DBTB INT Eric Williams r

my name is Eric Williams your curve data science at amada health I think it was being around people that see the potential of data especially in kind of the life sciences and in health I think there's a little bit of transition there at least in healthcare thinking about outcomes based pricing paper value rather than pay for service and the intersection of that in data science i think is particularly interesting and being around getting to talk to people that also are excited about that and the potential value there is really exciting I think data is cool and exciting because it's starting to be directed at really important outcomes so it healthcare is the obvious example I feel like we can take the same data science toolkit of analytics machine learning and experimentation has been optimized on you know traditionally generating a lot of ad revenue or click through rates or funnel conversions and replace those outcomes with health outcomes with the same tools and the same power and the same volume of data and I think what's really cool is that in with health systems it can kind of take a model like Netflix a where the more users Netflix get acquires watching their movies telling Netflix what they like the better prescriptions Netflix can give for movies that you may want to watch health care and hopefully advantage advanced advances in interoperability and data collection and data analysis we can get to a similar self learning system where input such as biology genomics and cancer histologies physical activity social dynamics can be tied directly to health outcomes in that sense the more patients that are treated the more we can learn about personalization of treatments for specific outcomes health outcomes I think that's what's most exciting the main insight is health care needs help and there's a ton of opportunities not as far as like personal data science opportunities which exists tube opportunities to really make make a difference and account for huge inefficiency gaps in the systems right now where the incentives are completely aligned the wrong direction and data has the opportunity to help align those and some incentives for real health outcomes I think that's the main thing I want to communicate I think to become a data scientist it would be getting exposure to as many different problems as you can whether it's a big big data or large data small data with exposure to analyses in different spectrums of data in types you are going to exposed to different statistical analysis the same stats that apply at you know particle physics level are very different than the ones that apply in the hospital it's good to get a view on that whole spectrum and the fidelity of the data changes to between those contexts the more exposure someone can give themselves whether it's your open data or different projects hackathons I think the more a data scientist can internalize that it's really the statistics underneath it rather than any particular latest and greatest machine learning algorithm or programming language to the underlying statistical intuition that you can translate from one setting to the next and become very perfectly two scientists you