AI Powered Recommendation Engine

Proudly Made in Tampa

Learn More About Hackathon

Original project was created in 45 hours during the Hack Hospitality in Tampa, FL.

Technical Details

Dig In

One of the sponsors, Sourcetoad, wrote a summary of our initial submission for the hackathon.

Why AI

Andrew Ng is famous for calling AI "the new electricity". We couldn't agree more. This technology has the potential to transform a broad range of industries, including hospitality.


We are five friends that came together in August 2017 to participate in the Hack Hospitality event in Tampa, FL. Our team was comprised of two front-end developers, a machine learning engineer, a back-end engineer and a designer. Notably, three team members were students of The Iron Yard. After the hackathon we have continued to expand the core technology.

AI Powered Recommendation Engine

We built a recommendation engine - powered by machine learning - that helps improve the quality and relevance of excursions that are suggested to cruise passengers. Since the hackathon, this core engine has been iterated on to provide a powerful technology to power recommendations from social media data using scalable machine learning technology.

Design & Front-End

The production ready front-end of the original application was a rapidly developed react SPA built by Keri Spencer and Mandy Jacobsen. Taylor Cox did all of the excellent visual design for the original submission and technology demo..

Data Collection

To collect enough data to train our initial model, Alex Spencer wrote and launched an entire Ruby on Rails application to collect survey data from participants from all around the world within the first 24 hours of the competition. He also wrote the data collection service that gathered information about existing cruise excursions (we were not provided this data during the competition).

Model Development

Mandy Jacobsen and Rob Venables worked together to create an novel way of converting items detected in photos to markers that could be fed into the machine learning models that powered explori.us. Mandy Jacobsen intelligently structured marker data to work with our limited data to increase the predictive power. We've since refined this design significantly.

Recommendation Engine

Rob Venables designed and built the machine learning recommendation engine that powered the excursion suggestions. The final solution chosen for the contest was four independently trained models that tried to predict excursion preferences for a given user for duration, price, activity level, and category.

Our Team

Rob Venables

Data Scientist / Full Stack Developer

Mandy Jacobsen

Full Stack Developer

Alex Spencer

Full Stack Developer

Keri Spencer

Full Stack Developer

Taylor Cox

Designer / Front-End Developer

Contact Us

Interested in learning more about this solution or want to talk about integrating the technology into your product? Send us a message - we'd love to chat.

[email protected]

Copyright © EXPLORI.US