Innovation Design Engineering (MA/MSc)
Bohao Zhang
Bohao is an interdisciplinary designer who has been working in the IT industry. He has a strong interest in mobility, urban design, AI and information flow. He received inspiration from many of ByteDance's products. Bohao's intent of the project is to automatise everyone's travel plan based upon big data, AI and users' preferences.
Why do travellers need such a service? Because of the difficulty of finding a good place from excessive amounts of information online, many famous landmarks are crowded. Foreign metro can be like a maze. Fragmented time can be unused. When we finally planned out everything, the trip can be a didactic task to finish. The serendipity and joy on the trip are lost. All these problems can be solved if we could have a professional guide who takes us on an adventure based upon a thoughtful understanding of our needs.
Kaleido
Our solution is to rethink the role of a navigation app in traveling. We believe it is not only a tool that shows the user the fastest way from point A to B but recreates an adventure with serendipity. Users input their destination and a type of adventure they like. Our algorithm creates a fully featured itinerary that fulfills their needs based upon their personalities. Using Machine Learning, we can generate truly personalised adventures at the click of a button Mapbox and Yelp API. We need at least 70 initial users to feed the algorithm so it can make basic predictions including planning sightseeing and going to a restaurant for our users. Before our first release, Kaleido will obtain the necessary needs of users and quantify them in a tangible way in our algorithm.
The core of our algorithm is to understand and predict people’s familiar needs in an unfamiliar situation. We are proceeding with user testing and insight analysis. We have got so many factors that could affect the prediction of serendipitous travel experience. They could be gender, safety, previous ratings, stamina, weather, and personalities. However, in this early stage, we start by suggesting restaurants and sightseeing locations. I anticipate the start of Kaleido’s service in London for three months and expand the service in major European cities by the end of 2021.
The core of our algorithm is to understand and predict people’s familiar needs in an unfamiliar situation. We are proceeding with user testing and insight analysis. We have got so many factors that could affect the prediction of serendipitous travel experience. They could be gender, safety, previous ratings, stamina, weather, and personalities. However, in this early stage, we start by suggesting restaurants and sightseeing locations. I anticipate the start of Kaleido’s service in London for three months and expand the service in major European cities by the end of 2021.
Existing Issue 1: Overcrowding on top destinations — The more famous attractions get more tourists. Top 10 countries account for approximately 48% of inbound trips, and the remaining 200 countries and regions counts for one third of total inbound visitors.
Design Opportunity 1: Lead people away from traffic — By acquiring data from mapbox and amap, Kaleido’s POI recommendations lead people away from traffic.
Existing Issue 2: Too novel or too boring — Raymond Loewy, a French born American industrial designer, believed that consumers are torn between a curiosity about new things and a fear of anything too new. As a result, they gravitate to products that are bold, but instantaneously comprehensible. Loewy called his grand theory “Most Advanced Yet Acceptable”. The power of familiarity seems to be strongest when a person isn’t expecting it, which generates serendipity.
Design Opportunity 2: Serendipity = Most Familiarity + Some Novelty — There second algorithm rule is to discover serendipity, which is 80% Familiarity + 20% Novelty. There are many hidden gems of the city that we have never been, which should take about 20 percent. The rest of POIs are where we familiar.
Existing Issue 3: influence from comments & reviews — From the 25 user interviews, I have also found comments and pictures that really influence people’s decision to go to a point of interests.
Design Opportunity: User Generated Content — Users’ rating to the POI affects Kaleido’s recommendation in the future. AI Model will better connect you to your preferred POIs.
Composition System — A user can type in their needs to generate an itinerary. The user doesn't need to type a sentence with any specific requests, but may only type in a blur request. For instance, "find a serendipitous morning trip for me." The AI system may generate a tranquil walking itinerary in Holland Park, and perhaps recommend a cappuccino at a small coffee shop hidden at the corner of the street.
Select Interest & AI Planning — The user can select, reorder, add and delete the itinerary in the preview page based upon interests.
Comments & Preview — Users’ comments and reviews will influence the model and future suggested itineraries.
By the end of 2017, tourism represents 292 million jobs and 10.2 percent of global GDP, and the number will keep growing in the long run. If travellers spread out around the world evenly, the industry will be easier to absorb, and travellers will also have a better experience.