Hodo helps travelers plan their trips by utilizing AI to recommend personalized places and activities based on the user's travel personality. The project targets anyone between 18 and 65 and aims to reduce users' research time while recommending unique places and not just generic touristy stuff. The AI system learns from user feedback and alters recommendations accordingly, making trip planning more efficient and enjoyable.

Currently, the users are taking the following steps to travel:

Issues with the current steps:

1. Time-consuming: researching things to do takes approximately 3-4 hours. The planning takes some more time. If it is a group trip, planning also involves video calls and a few extra hours of talking it out and planning.

2. Generic: Even after all this research, the places you find are generic and touristy. You must do additional research to find the hidden gems or places perfect for you. 

3. Currently, no applications plan the trips for you—just many booking applications and budget-planning ones. Which is excellent, but this comes only after the planning stage.

• Expedia, Trip Advisor, Google Flights, Hopper, and Kayak use AI to help users get the best deals on flight and hotel bookings. 

• These applications also help with budget planning. 

• Trip Advisor and Expedia also have articles that help with things to do at your intended destination. 

• These applications use AI to make the best possible recommendations for flights and hotels using attributes like ratings and filters like the number of days, budget, destination, number of people, etc. 

• They also have things to do section. But this section lists “touristy” things the users can do at that particular destination. 

• This section is filter-based with options for kid-friendly and pet-friendly places, but it does not list out the hidden gems, nor does it cater to specific user personalities. The user must still look for sights manually, read the description, and decide whether they want to include this place.

Expedia

Quizzification is a more effective and fun tool to gather user information. Users prefer that significantly over the traditional form-based sign-up system.

• Buzzfeed quizzes are fun - 67% of people interviewed had, at some point, taken a BuzzFeed quiz for fun. 

• They found this an exciting way to get to know their personality. 

• They are also intrigued by how these quizzes work and what kind of mapping, process, or maths is done to get the results. 

• Headway - an app that provides users with ~ 15 minutes reads to help them set goals and teach good habits- is the perfect example. 

• The application has an onboarding quiz that comprises this or that - you choose what book “sparks joy”; there are also multiple choice questions about your preferences. 

• The users answer these questions and have fun while doing it.

This is when we decided to narrow down the personas we wanted to design for

• People plan trips differently - 60% have loosely designed itineraries with room for spontaneity—27% like to plan things down to the last detail. There is no room for spontaneous plans and last-minute changes. 13% of people are spontaneous and like to go with the flow. They want to keep a handy “things to do” list.

• People have mixed feelings about getting the trip planned through travel agencies - 14% of people prefer to get their trips planned by travel agents so that everything is taken care of. They find the agents trustworthy and feel they know better. The agents also make bookings for these people and get them the best possible discounts. They think this is a more time-saving and stress-free option.

• 86% of people prefer planning their trips because it is fun, and anticipation and planning are also part of the experience. Agents give itineraries that could be more generic and more tightly planned. These are the group of people that like a little spontaneity and prefer to explore the place less explored.

This was a 3-month sprint where we had to research, design, and develop (code) a prototype (including the UI and AI algorithms) and write a thesis. So, for this iteration, 

• We decided to work with just a few features - Sign up, personality quiz, sightseeing recommendations, save wishlist, search more places, add new locations to saved wishlist, and improve recommendations based on new additions.

• We used a small data set that we made ourselves containing places and things to do in Italy. 

• We also created a small set of personalities to work with that were overlapping in nature and contained characteristics of some of the 12 personalities combined so that sorting would be easy. The following are the sample personalities that we came up with guided by the paper by Gretzel et al. :

Based on all the information gathered so far, we created the user flow diagram and sketched a draft of what the flow would look like.

After we got the user flow down, we designed our wireframes.

The Personality Quiz

Soon after we designed the wireframes, we created the mid-level prototype. The plan was to have a basic clickable Figma prototype that the users could use on their phones and try out.

We narrowed the potential ideas and designed a mid-level prototype based on the flow identified in the service design blueprints. We created a mid-fidelity prototype and moved on to evaluate that.

We put eight people (students and staff, ages 22-50) from our campus (California State University, East Bay). We asked them to do a few tasks using our phones, and each session lasted about one hour. The tasks were as follows:

• Sign up and create a user profile.

• Take the quiz and find your personality

• Retake quiz

• Add your destination and get your itinerary recommendation.

• Save an itinerary

• Edit the itinerary.

• Search for a place that you were not recommended

• Add that place to your saved itinerary

• Create your itinerary

• Sign out.

Our goals for the test were:

1. Test the accuracy of the personality quiz.

2. Test the discoverability of the features such as trip recommendations, saved trips, and search places.

3. Learn about users’ travel habits and behaviors as they navigate the application and complete the task.

4. To conduct a feedback session to understand what, according to the users, were the challenges and opportunities that they came across while using the application.

The users loved the Buzzfeed-like profiling for sign-up and had fun retaking the quiz to see if the answers affected the result for personality. But they had a significant issue with the design system. They did not like the current design of the mid-level prototype, and hence we redesigned the application entirely. We also had three key findings and improved upon those flows.

Key Finding #1: The onboarding screen needed to give more information about the application. 

In the test, some mentioned that the information the onboarding screen provided needed to be more precise and give clarity about the quiz and user profiling and any information about what the application does.

Solution #1:A 4-screen onboarding 

Therefore we added a 4-screen onboarding that lets the users know how the entire app works.

Onboarding Flow

Quiz Flow

This project is just the tip of the iceberg. We still need to do a lot of research on the personalities. AI needs to learn a lot and revise and update a lot more. The quiz must also be more indirect and fun or have fewer words and more images. The current questions are direct and leading, so the users can come in with prejudice and answer questions with options that give them a specific result. The users could also feel that they are being judged for the outcome and might try to get a personalized result that feels good rather than accurate.

We are currently working with a small dataset and would like to know if there will be any new challenges the algorithm will face while working with a larger dataset. Will the algorithm need to be modified?

We also need to add more features - An itinerary that does a day-to-day plan for the users, adding geolocation where you can plan a route for the trip, alerts and alternate recommendations for when a place is closed, or the weather changes things, or specific locations are only open in certain seasons.

This project personally was a great lesson for me. I was able to lead a team of three people and as the Product Manager of the team along with being the UX Researcher, I was able to create a product development strategy, communicate with the stakeholder - which was our thesis panel - and plan the pitch meetings, work with the UI designer and developer both as a UX researcher and a PM. I was also responsible for product related documentation. I also learnt how to plan small sprints with features delivered according to priority. This project not only made me level up as a UX designer and researcher, but also introduced me to product management.