From the Nike swoosh to the Starbucks siren, logos are one of the most instantly recognizable parts of any brand. Many logos have become icons unto themselves, and the public feels deeply attached to them – so much so, that companies often face significant backlash when they try to make a change. The task of designing a logo has historically been viewed as an art more than a science, but a new paper by Wharton marketing professor Ryan Dew proposes a more data-driven approach. He recently spoke with Knowlege@Wharton about his findings, which are outlined in the paper, “ Letting Logos Speak: A Machine Learning Approach to Data-driven Logo Design .” The paper was co-authored with Asim Ansari and Olivier Toubia, both professors at Columbia Business School. An edited transcript of the conversation follows. Knowledge at Wharton: Why are logos so important for brands? Ryan Dew: Logos are kind of the visual figurehead of any brand.…