TeamMopped

FrontEnd_1
FrontEnd_2

Inspiration

With the rising popularity of games and gamification becoming more and more part of our lives many opportunities rise to bring gamification to more and more people. We found our initial idea as true Slavs in the bars. They are called stock bars because prices in these pubs change based on customers' demands. This sparked idea, of how we could bring this idea to more parts of our everyday lives and we created "Stockerize".

What it does

Stockerize allows vendors and sellers to dynamically change their prizes of merchandise or products based on customers' demand. On other hand, this creates new and promising sales for customers to pick hidden gems from their favourite vendors. Last but not least it is built with state of an art recommendation engine in mind, which feeds users with prediction sales tailored for them.

How we built it

With our skillset which perfectly overlaps, we created a simple modern FrontEnd (FE) application in React.js. This FE app communicated with the BackEnd (BE) application - which serves the role of our main server in architecture. The server is built with the Python Django framework. Our main data storage is MariaDB which is deployed in the cloud same as our Django server. Both of these applications are in the Microsoft Azure cloud. For more convenient use we also created CI/DI for all our applications with one button deploy. The recommendation engine is built on rock-solid machine learning principles of cosine similarity - especially KD Trees, but we have more advanced ML models down in the pipeline.

Challenges we ran into

There were several challenges we tackled. Here are some examples.

As all participants of HackKosice 2022 we had problems with internet connectivity, so coding without google was extra hard these days.
At the start we tried to build a deep neural network for recommendations but after we run into long training times we choose to switch to a more machine learning-oriented principle.
When we headed a hackathon with certain principles in the back of our minds, one of them was to build the most scalable app. So docker was right no brainer choice. After struggling with FastAPI, we have decided to scratch FastAPI and migrate all our code to the Django framework. But creating tailored docker for our application was a little more problematic than we thought.
When we selected our challenge we had no idea about e-commerce so trying to learn as much as possible in a short period of time was challenging especially the business side of e-commerce.

Accomplishments that we're proud of

Our most accomplished achievement is the APP. Everyone on the team is satisfied with the outcome and that we pushed our limits to the maximum.

This was our first time coding as a team and we consider our teamwork without any hiccups as a big accomplishment too.

What we learned

We learned a lot.

E-commerce from the business side is not an easy job to do. (Programming same)
We had zero knowledge about recommendation systems, through the hackathon we learned about their structure and how they work.
We learn how to sleep with open eyes or how to overcome the need for sleepiness.
Not giving up is a viable option and its in our nature.

also, MEGADYNAMICKE is a new buzzword you see here at first.