BI Trends

DeveloperWeek New York 2020 Hackathon [Guest Post]

By Akhil John
Share on LinkedIn Tweet about this on Twitter Share on Facebook

This is a guest post by Akhil John, Software Developer at TrophyTracks.

DeveloperWeek NYC hosted a virtual Hackathon this month with a prompt of “Rethinking Travel.” Read on to hear from a member of the winning team for the Logi-sponsored hackathon who created a solution for traveling in a post-COVID-19 world.

login

How It Works
Travelogue allows users to take and post pictures or videos of their trip, or events around their area. The application stores all the photos or videos the user has taken with their geo-tags and populates the feed with recent posts from around the world. When browsing the feed, users can see other users’ posts, while we provide potential flights and hotel details. Additionally, the user can explore activities and tours in the area.

Building the App

IMG 20201209 072447938
I started by building the Travelogue app with Google’s Flutter framework. This helped us build our user profile, store user data, and store user images all in one place. Building with the Flutter framework minimized the development time as it integrated flawlessly with Firebase.

I built the login interface that allowed our users to sign in through Google and proceeded to start building the user profile. I then started building the homepage that generates the feed of all user posts by location. Since we lacked user data, I wrote a Python script that populated our database with images and meta data of different popular cities. Time was of the essence, as we had yet to look at the different API for the sponsor challenges.

Then I looked at Logi Composer.  Signing up for the free trial was simple, and I was able to understand how you connect to and prepare data without the need for heavy data modeling or coding. One challenge that we ran into with Logi Composer is since we were using the trial version, we were limited in the sources that we could use. However, this was easily resolved because we had the CSV file that we used to generate the user data, which we used as the source.

Likes added by new posts over time

Uploading our user dataset to Logi Composer, we were able to generate some very interesting visual graphs to better understand our user data. We started off simple with a comparison of user posts over time and its relation to the total number of likes on the app

Phototag Usage

Next, we started playing with Logi Composer’s more creative features and created a word cloud of the most used tags on photos. The words size represents the number of times the word was used and the color indicates the amount of likes.

User Sentiment of Cities

This visual shows the volume of posts compared to the relative number of likes. It highlights user sentiment, for example, even though there are few posts in Toronto Canada, they are well liked.

We then implemented the Amadeus API (as part of the Hackathon) to pull all relevant flight and hotel information for a potential trip. I implemented a REST API call to get the flight and hotel information for each post, so users can get the travel information for that specific location.

The final couple of steps that we had to take were recording the video for our demo and uploading the application to the Google Play Store. Here’s our one-take video:

 

What’s next?
Looking ahead (beyond the five-day hackathon), we would be increasingly focused on the user experience. The main targets would be transitioning Logi Composer from just an administrative tool to being integrated into the app for users. This integration could directly support users finding new areas to travel to, as well as feeling more secure in their travel choices with a Logi Composer generated view of COVID-19 affected areas.

The user base will also expand as Logi Composer visuals are enabled, by using its functionality we hope to allow users to see analyzed sentiment from user’s posts. By making the app as general purpose and inclusive as we have, we hope to negate the typical rating system where individuals only rate extremely positively or negatively. This will give users a fuller understanding of the subject and may help showcase different tastes without discriminating against any user groups.

This hackathon, although virtual, was a fantastic experience. There were a range of speakers that discussed various topics in technology which was not only informative but also thought provoking. The hackathon also included helpful seminars that showed how to integrate both Logi Analytics and the Amadeus API in our application. I want to take a moment to thank both the sponsors and DeveloperWeek for hosting the hackathon. Looking forward to the next one!

Originally published January 4, 2021; updated on January 5th, 2021

About the Author

Akhil John is a Software Developer at TrophyTracks. He has years of experience in application development and machine learning. He is looking for opportunities to grow in the field of software development and embedded systems.