Find the best place to move to with our machine learning algorithm!

What it does

Based on an analysis of multiple sources of data, we have designed an algorithm that can take into account the preferences of the user, the predicted happiness of the person being in various countries that optimize for each question and finally choose where a person would be most satisfied in the world.

How we built it

We built up our frontend with HTML, the backend was designed with node.js and javascript. The data mining, merging, wrangling, extraction and parsing was done in R before being passed to the javascript code to handle the finer details. Our machine learning algorithm takes into account the priorities of the person as the initial weights and the global ranking index as the node value before computing the maximized sum to select the country with the least compromises.

Challenges we ran into

Our original plan was to use Tensorflow.js to create the program, however, we ran into difficulties getting set up with it and ensuring that models built would be transferred properly. We switched over to brain.js that had a better/simpler library for use. However, its constraint was that it could only handle numeric values. Thus, we settled on building our own custom classifier to choose which country would be best taking into account all factors

Accomplishments that we're proud of

We are proud of building our first complete web application that can take in user input, compute useful information, run it through a model to find an optimal solution and present it to the user

What we learned

We learned how to apply machine learning in a web development environment and debug challenges when writing for node.js vs the browser.

What's next for BonVoyage

Being able to customize the survey with better questions that can take into account user biases.

Built With

Share this project: