Most Data Science initiatives & projects fail. By some estimates, when judged by practical or actionable outcomes, less than 5% of all ML, AI, Analytics and Deep Learning work may be considered successful. For a profession which prides itself in ‘knowing numbers’, this is one number, which is often disregarded by the Data Science fraternity. One can only imagine the quantum of resources, time, effort, investment, and planning, poised to go down as failures. Based on our on-ground experience, interactions with business leaders and working with data teams, it is clear that most of these failures are attributable to three basic flaws: 1. Assuming ML & AI solutions are plug & play 2. Not linking data science outcomes to business goals 3. Building a team of more dreamers than doers This article by the ‘Data Doc’ talks about some of the soft issues which are connected with the above points and is an essential read for anyone looking at setting up, hiring or rebuilding their Data Science teams. Data may be the new oil… but we seem to be building inefficient refineries.
View this Post on Linkedin