Clash Royale Deck Analyzer

drag and drop deck builder with blank deck
deck builder with example deck
searching for cards based on name
searching with multiple values
searching on even more values
building deck by inputting card names by text
text based analysis of the deck for colour blind users
visual analysis of the deck showing archetype and composition

Inspiration

Inspired by the Grafana dashboard, we sought to create a unique version tailored to Clash Royale, a popular mobile game. Noticing a gap in deck analysis tools, we decided to build our own solution that not only analyzes a deck but also provides a wealth of information to help players optimize their strategy and improve their chances of winning games.

What it does

Our application allows users to easily input their Clash Royale deck either through a graphical interface, by pasting a deck link, or by manually typing in the cards. The app then provides a detailed analysis of the deck, displaying key stats such as the 4-card cycle, average elixir cost, and deck archetype, all backed by data from the top 100 players in the game. The analysis is presented visually through engaging graphs and pie charts, helping users understand their deck's strengths and weaknesses while offering strategic insights on the best ways to play and maximize their chances of winning.

How we built it

we found the 100 most commonly used deck from an API and manually classified them as the various archetypes this was then used as training data to create our classifier that provided a confidence rating for each archetype. then we created the GUI based from the in-game version to create a deck to perform the analysis on.

Challenges we ran into

lack of existing training data which we solved by creating it our own. issues with bugs created by gen-AI that had to be solved by manual debugging, we also began to run into a point in the project where the AI could no longer make successful modification to the code and development had to proceed manually.

Accomplishments that we're proud of

We are incredibly proud of our achievement in training a highly accurate model using just 100 decks. We’re also thrilled with the features we’ve implemented, including a user-friendly graphical interface and text-based input methods, which together provide a seamless and versatile experience for deck analysis and strategy optimization.

What we learned

how to integrate AI into our workflow, how to gather and create training data for a neural network, how to train a neural network, how to make a image based GUI with search function. How to make and post a YouTube video.

What's next for Clash Royale Deck Simulator

In the future, we plan to integrate an AI assistant to further assist players in mastering their deck, offering personalized tips and in-depth guides for each archetype. Additionally, we aim to enhance the user interface, making it more modern and visually appealing. We also plan to introduce features such as EVO variants for cards and other small improvements to make the app more comprehensive and user-friendly. Our goal is for the app to be web-based, allowing users to access it seamlessly from any device.