A recipe for success in exploiting machine learning and data science
An often-quoted promise of big data and data science used to be: invest in databases and algorithms to process data and you will gain new insights, even reducing the need for subject matter experts. The seduction of a life of leisure seemed within our grasp!
While this is commendable in its eagerness to encourage better exploitation of data, the reality is very different – as explained by David Hand of Imperial College London and senior technical manager Malcolm Moore. In addition, JMP senior systems engineer Emmanuel Romeu provides a live demo.
Effective data science needs:
• Communication and teamwork
• Clarity about what you want to know
• A firm grasp of the data
• Understanding of the quality of the data
• The ability to extract information from data
Much of this is beyond the current capabilities of databases and algorithms. As such, subject matter knowledge is needed at all stages of extracting helpful knowledge from data.
Put another way, data science and big data need subject matter experts to succeed. Companies get the most value from data by enabling experts to apply statistical and data science methods, resulting in better decision making with increased certainty.
