.@HarvardBiz: A #DataScience #ValueChain such as Research Scientist, Data Engineer, Causal Inference Specialist etc…coordinated by a Product Manager, is a useful conceptualization. But since there are no blueprints to follow & ‘you learn as you go’...https://t.co/rs8qPwX6z3
...and multiple iterations are involved, excessive specialization inordinately increases coordination costs & wait times. So, start w generalists - '#FullStack #DataScientist', but modify if cost of error is high, or eg yottabytes of data! via @digitalarati @ericcolson @netflix