Inspiration

How many times you were standing ages in the queue while only wanting to buy one bun? How many times have you wished that you knew beforehand, what would await you in the shop? Q(ueue)Wise aspire to solve this!

What it does

QWise is an application in which the main purpose is to help YOU waste less time queueing and make your life easier. You don't even have to type your questions, you can just voice them! What's more, the application supports the Polish language. Just say you want to shop. The assistant will tell you what is the best time to visit your nearest shop or which shop in your surrounding has the shortest waiting time right now! It all depends on what you prefer. You can also specify the exact time and location you are interested to shop in and set your default home location or another place you usually like to shop in.
What is more, QWise understands the new pain of Polish people, this is Sundays without sales! Not only the app will remind you about it but also will include this information in response to your questions.

How we built it

Our project is composed of two components: the Google Assistant fulfillment and Java backend service.

The Google Assistant part is the best summarized by the linked document we made.

When it comes to the backend service, initially we intended to get information about payments made with VisaCard in a given time and place. The VisaApi (Visa Queue Insights) not only would we get the information about Visa Payments, but also estimation about MasterCard users and paper money enthusiasts. Unfortunately, as of right now the application is only available in the USA. So, we used a web-crawler to get the data from the websites about the approximate density of people in different hours on different days of the week. Then we populated our database with this information. Then we build a server which sends all necessary data to the Assistant.

Challenges we ran into

First and the biggest challenge was that the Visa Appi we wanted to build our application on, is right now available only in the United States. As we wanted o incorporate the newly released Polish version of the GoogleAssistant. So we had to get the data ourselves, which was really time-consuming as we wanted to have a reliable, 'life' examples.

The lack of sleep didn't help either, but that just how it is on hackathons :D

Accomplishments that we're proud of

First of all, we managed to connect all our work into one, working application.

We are proud that we got to put our hands on the data that cannot be easily obtained. As we wanted to have reliable database entries, we build a web-crawler with the Beautiful Soup, that we used to scrape from popularity graphs from Google Maps.

Whatismore, the Assistant application has a natural dialog flow, witch help with the user experience. And we connected with GoogleMaps to navigate to the chosen shop.

What we learned

Most of our team members have experience in working with GoogleAssistan, but it was our first time creating a backend application with which the Assistant connects and gets data from. It was a real learning opportunity which was not only really condensed in time but also intense, and I mean really intense.

We also learned how to scrape data from websites using Python and refreshed our knowledge about Heroku deployments.

And most of all, we developed our teamwork and time management skills.

What's next for QWise

What we want to do next:

  • Add different types of shops and places, from which the user can choose from. Starting on clothes through the board games ending on the fishing baits (right now we have the grocery's as a default option)
  • Add different priorities with which the shops will be selected with the possibility to add custom filters
  • Taking into consideration not only days of the week but also the months, as Friday evening on July differs from Friday evening on February
  • Making predictions about exact dates based on the data from last years
  • Connect to VisaAPi and Linkyfi Api as soon as it will be available for us

Built With

Share this project:

Updates