Feeltrip

Inspiration

We chose travel within the entertainment path because there is a constant increasing demand for travel, and studies have shown that millennials care more about the experience than material goods. However, there are not a lot of sources to get customized travel information. All sites have bountiful information about attractions, but traveler has to search everything to plan his travel. As a result, we suggest customized service for them to make their planning more convenient and satisfying. In this way, the traveler can save time and focus on more of what they want to do during their vacation as opposed to worrying about the details that they need to find individually. The idea is to offer a platform to find attractions that are specifically catered to one's interests.

What it does

The user enters a destination city and chooses between two prompted images based on their preference of one choice over the other. For example, if the user chooses Washington, DC, pictures depicting city view vs nature scenery would be prompted, and if the user chooses nature, mountains vs water (ocean, lake, pond, etc) would be prompted. In this way, the user's preferences will be sorted out and local attractions will be found so that the user can pick and choose what they would like to experience. A brief description of the attraction will also be provided for the user. On the attached PDF slides, there is an example of how this app would work with the sample UX slides.

How we built it

The primary dataset has different city attractions and includes information such as subcategories, location, description & images (URL), and reviews. The subcategories of the dataset are used to prompt the user about choices (i.e. things like city vs nature would be subcategories). By utilizing Google vision API and web/image crawling, we can get code keywords to find images of attractions pertinent to what the user has previously chosen. As such, a list of attractions that would potentially interest the user would be created, from which a user can choose whichever one they are the most interested in. A description and location of the attraction is then given.

Challenges we ran into

Setting/usage of API Although we couldn't fully develop our idea with the Google Vision API, we used all the resources available (YouTube videos, "Getting Started" directions on Google Cloud homepage).
Web page crawling: A difficulty we found with web page crawling was that the information online for descriptions/reviews of an attraction was in all different types of formats so one uniform code wouldn't work for all applications. However, we were able to do image crawling, and the code for that is available in the attached PDF slides.
Finding enough appropriate datasets: We found a dataset for London but the more data we found, the more disorganized they were and it was difficult to find datasets that had all the same categories for many different cities. Additionally, there was not a lot of sufficient datasets for this application.

Accomplishments that we're proud of

Practicality of idea which might impact people's lives to make traveling more convenient/customized
Social impact: ability to have a more tailored experience that is similar to asking a local about things to do based on individual interests
User impact: the app aims to gather all available data for the user to get the top suggestions for each individual. It's easy for the user to narrow down information and have all of it in one place. There are not many apps that give suggestions based on one specific user's preferences.
Tourism impact: advocate tourism and this could be used in locations that thrive from tourism to attract more visitors. This would have a greater overall economic effect. It could also be used to bring interest to places that would benefit from more tourism.
UX - the app design simple to use, easy to recognize, straightforward and very intuitive.

What we learned

Setting up API
Crawling via Node.js
UX designing tool
Image crawling using Python
Various Google services

What's next for Feeltrip

Indicate popular hours, crowded times
Recommendation of food nearby
More information about attraction such as transportation, festivals in the area
By utilizing Google Weather, app can indicate what dates are most appropriate for traveling.
Based on Review data on our primary dataset, app can indicate reviews of tourism spots from customers
If this app could be linked to Facebook or Instagram, this could also recommend some spots for users based on the previous "likes" of the users.
Also, the app could recommend best photo spot for people by connecting image to geographical locations.
In the future, the app could also recommend users of more specific options (i.e. panoramic views, best times to go to certain places)