Bringing quality healthcare home

Where and how to launch 'Bobocabins' in the United States

Rethinking urban mobility in Europe

Action Learning


Developing a web app dashboard to identify children's quotes for greeting cards


LittleHoots is a Kansas City, MO startup founded in 2013 to provide parents with a way to preserve memories; their app makes it easy to capture, style, and archive children’s quotes. Seeking to leverage this content for merchandising, the company asked MIT Sloan’s Analytics Lab (A-Lab) for help automating the process of finding marketable quotes from within its database of more than 500,000 quotes, conversations, and stories. The A-Lab team addressed this challenge by developing a clustering model to extract keywords based on the frequency with which certain terms appeared in the children’s quotes. To address a scarcity of labeled examples, the students then used easy data augmentation techniques to create additional examples. This work improved the accuracy of the team’s natural language processing (NLP) tool. Ultimately, the students delivered a web app dashboard that LittleHoots can use to explore captured quotes by topic and identify appropriate quotes and conversations for consumer products—specifically greeting cards.

“The biggest challenge was dealing with the limited amount of labeled data,” says Jorge Castillo, MBA ’21, who worked on the project with Alex Adamczyk, MBA ’21, and Chiayi Kung, MFin ’21. “We had to think out of the box to come up with creative ways to create more data and label it. This was critical to properly train our models and improve performance.”

Lacey Ellis, founder and CEO of LittleHoots, says she loved the team’s model and plans to use it to grow the company’s merchandising business. “I am honored and thrilled to have worked with such a brilliant team of students at MIT! They built a sophisticated NLP tool that will unlock a whole new revenue stream by enabling LittleHoots to uncover hidden gems of pint-sized perspective for greeting cards and books,” she says.

The A-Lab project was also a great learning experience for students, according to Castillo. “Not having a lot of coding or data science background myself, A-Lab really stretched me!” he says. “With the help from my teammates and faculty mentor, I was able to significantly improve my analytics skills and couldn’t feel more proud of the end result.”