23 Questions Successful Data Science Project Asks Before Embarking

Why do most Data Science projects fail to achieve ROI? How do conduct Design Thinking workshop for Data Science projects?

Kunal Sawarkar
IBM Data Science in Practice
7 min readMay 23, 2019

--

Bad Reputation?

The joke about data scientist is that, “ A data scientist is someone who is very successful at solving the wrong problem.” This is normally aimed at enterprise data science projects which don’t always deliver business value that justifies the investment in terms of time & infrastructure cost. In fact, as this Harvard Business Review report points out, most organizations struggle to deliver positive ROI on the numerous projects they begin and have very high failure rates. Normally, projects can deliver a very accurate model and still be a commercial disaster by being unable to achieve the desired adoption by business stakeholders. What explains this apparent paradox ? Why can’t we get that kumbaya moment that everyone hoped for when they started on the AI journey?

Trifecta of Mistakes

  1. Rabbit-hole- The project starts with one problem and then wanders into others as it moves along. Objectives gets revised to accommodate either available data or newer demands. Soon the project finds itself inside a rabbit-hole since it did…

--

--

Kunal Sawarkar
IBM Data Science in Practice

Distinguished Engg- Gen AI & Chief Data Scientist@IBM. Angel Investor. Author #RockClimbing #Harvard. “We are all just stories in the end, just make a good one"