Spring 2018 Winning Project: Wander

Revolutionizing how travelers discover meaningful experiences by providing memory-driven recommendations that connect familiar emotions with new places and cultures.

--

Team

Martha Czernuszenko, Strategy Lead
Su Fang, User Research Lead
Peter Shortino, Technology Lead
Taylor Nelson, Design Lead

Research Insight
Final presentation. Advanced Design for AI course at The University of Texas, April 25, 2018.

Background

Through our research into what people consider when choosing experiences during a trip, we discovered that travelers often rely on fond memories and their own cultural touch points to find connections of comfort and familiarity when exploring new places.

With this insight, we were inspired to create an app that would allow us to go beyond basic recommendation systems, like what you see on Amazon and Netflix, and offer users a way to “discover their new fondest memory” by understanding their values and motivations.

The name Wander comes from two sources: a sense of wonder you get when reliving old memories or exploring new places, and when you wander around a destination trying to find something to do. We wanted a name that was reflective of the user-controlled experience behind our app and how it was the user and their memories that were the driving force behind them creating something new. Not some professional, paid traveler or guide.

How it works

Wander uses posts, check-ins, likes, images, and videos available across online services and applications to understand who you are and the experiences you enjoy.

By cataloging patterns, entities, intents and sentiments in your data, it can analyze your current surroundings and make suggestions for new experiences that have similarity to memories that had the most meaning for you in the past.

For example, a traveler of Polish descent in China might be given a recommendation for a local dumpling stand because of their similarities to his grandmother’s Polish pierogies and the fact that it’s a street stand, like the food trucks he likes to frequent with friends back home.

Wander does not rely on you spending hours browsing the internet, asking Facebook for recommendations, or consulting those hefty (and often expensive) professionally written travel guides. Instead, it searches through your history of check-ins, photographs, posts, etc. to formulate suggestions based on things from your footprint it thinks you would like.

In our highlighted use-case of Sam (see the video above), we acknowledge that Sam’s memories are tied to his social media, and these memories can bring him comfort, joy, and nostalgia to serve as a context to understand his novel travel experiences. Sam’s digital media footprint is not only a representation of his potential travel interests, but also of near-and-dear memories, with significant individuals in his life, such as his close friends in Austin and his Polish grandmother.

In acknowledging this fact, we as a team recognize the significance of these posts, not simply as fleeting musings that express surface-level personality traits, but as emblems of Sam’s relationships, emotions, and memories.

Model

In the wake of recent data security events, one major question that the tech and social media industries are facing is, how should we conceptualize users’ data?

In creating Wander, our team’s stance is that user-generated media should not be simply conceptualized as individualized data points which form overarching patterns, but that many of these media posts already tell stories on their own.

They already carry meaning to the user just by virtue of being posted and shared. Creating experiences based on these stories not only provides more meaningful, personalized experiences for the user, but also shows a respect for user-generated media and the stories that they have to tell.

Privacy

Due to rising concerns about data privacy and corporate profits from personalized user profiles, we think a different approach is necessary.

Data privacy quickly became our number one concern in framing Wander’s business strategy in a way that could be profitable while respecting the privacy of its users.

Utilizing a blockchain encryption method, Wander guarantees the privacy of our users and their ability to opt-in and out of sharing at any time. Unlike social media sites like Facebook and Google which sell your private data for their own profits, Wander only sells bulk generalized data. Although not as specific, it would provide insights into how users move, where they go, what they rate, etc. — just not the specific user it is attached to.

This approach helps secure user confidence in Wander’s recommendations. In turn, they would be rewarded each time they make a selection, review, or search for a place with a series of tokens that can be redeemed for rewards like coupons, gift cards, ad-free browsing, etc.

Impact

Although we’ve framed our concept in the context of a travel app, the idea of using data and machine learning to understand memories that, in turn, provide insight around a user’s context, values and motivations is far more powerful than any recommendation system on the market today. Memories are our most emotional and sacred touch points.

A system that can understand what holds meaning for us and use that knowledge to help us connect with unfamiliar environments could be applied to any industry, domain, or perhaps most importantly — new cultural experience.

--

--

Jennifer Aue
Advanced Design for Artificial Intelligence

AI design leader + educator | Former IBM Watson + frog | Podcast host of AI Zen with Andrew and Jen + Undesign the Grind