DBTB INT Eric Williams r

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