Easy Tour

Planning a trip has never been easier because EasyTour's AI picks the right mood for your trip using computer vision!

Comment

Inspiration

Being one of the BUSIEST cities in the entire state of California, our vision was to make its numerous amenities and locations for accessible for ANYONE who happens to be driving aimlessly around Downtown wondering why nobody is giving them a tour of this great city! With our unique and completely original machine learning model- using the user's perception of color to train our model and allow the AI to recognize a spectrum of emotions- we were able to bring out the best from Google's Cloud Vision API (specifically image annotation and properties)! LAUNCH HERE

What it does

Easy Tour saves you the many disappointments of Tourism brought on by busy distractions such as traffic, poor planning, and worst of all, the possibility that you simply "weren't in the right mood" for the trip.

How we built it

Uses computer vision, trained on user's spectral cognitive perceptions, to categorize MOODS of travel, such as energizing, calm, exhilarating, melancholy, pleasant, and more. Photo input is fed from Flickr API, which provides GEO coordinates of a photo and allows the AI to categorize locations based on mood.

Step 1: Train your model by entering the place you want to travel, number of pitstops, and personal color moods.

STEP 1

Step2: Select your MOOD of travel (HA! Get it?)

STEP 2

Step3: Watch the magic happen

STEP 3

Challenges we ran into

Many API endpoints, and a steep learning curve for each one, especially asynchronous vs. synchronous calls to flickr API, and Google Cloud Vision API.

Accomplishments that we're proud of

We learned REALLY rigorous backend and set ourselves up to building a real professional project that we couldn't be more proud of! Even we surprised ourselves with the product we created, it was the first time we were able to leverage very high end technology to creating something nuanced and personal (no plagiarism involved).

What we learned

We learned to manage our time effectively, what makes a good workflow for this type of project, and what it takes to make a great product in a limited scope of time.

What's next for Easy Tour

What's next? More users, more predicting, and more fun! We want more people to try out our newly thought of functionalities and really see if what we made can compete up there with the pros! So try it out and let us know!

Built With

Share this project:

Updates

posted an update —

Example of our AI Spectral Model Code:

// Read JSON file to get city name and search flickr database for matching groups app.get('/moods', (req, res) => { var modelData = fs.readFileSync('../public/photoMooder/customModel.json'); var cityName = JSON.parse(modelData)['City']; var latitude = JSON.parse(modelData)['Lat']; var longitude = JSON.parse(modelData)['Lng']; var redEmo = JSON.parse(modelData)['Red']; var yellowEmo = JSON.parse(modelData)['Yellow']; var greenEmo = JSON.parse(modelData)['Green']; var cyanEmo = JSON.parse(modelData)['Cyan']; var blueEmo = JSON.parse(modelData)['Blue']; var magentaEmo = JSON.parse(modelData)['Magenta']; var whiteEmo = JSON.parse(modelData)['White']; var blackEmo = JSON.parse(modelData)['Black'];

... ETC ...

Log in or sign up for Devpost to join the conversation.