Blitz Review

Inspiration

As an avid League of Legends player, I was always curious about my performance patterns and wanted a more engaging way to review my gameplay beyond the basic match history. I was inspired to create something that would make year-end gaming reviews as exciting as Spotify Wrapped, but tailored specifically for League players. The idea was to transform raw match data into meaningful insights that could help players understand their strengths, identify areas for improvement, and celebrate their gaming journey.

What it does

Blitz Review is a comprehensive League of Legends performance analysis tool that generates personalized year-end summaries for players. It fetches match data from the Riot Games API and presents detailed statistics including champion mastery, role efficiency, performance trends, and fun facts like total creeps killed and favorite items. The application uses AI-powered analysis to provide structured insights across five key categories: overall performance assessment, champion mastery analysis, role efficiency evaluation, areas for improvement, and strengths to build upon.

How we built it

The project is built using AWS serverless architecture with a Lambda function handling the backend logic and API integrations. I used the Riot Games API to fetch player data, match history, and detailed match statistics, then processed this data to calculate comprehensive performance metrics. The frontend is a responsive web application that displays the analysis in an engaging, card-based layout with distinct visual sections. For the AI analysis component, I integrated Amazon Bedrock with Nova Micro to generate personalized insights based on the processed match data.

Challenges we ran into

One of the biggest challenges was handling the complex nested structure of Riot's match data and ensuring proper error handling when participant data wasn't available in the expected format. I also faced difficulties with AWS Bedrock permissions and model parameter formatting when switching between different AI models like Nova Micro and Claude Haiku. Rate limiting from the Riot API required careful request management, and parsing the AI-generated analysis into distinct visual sections proved more complex than initially anticipated.

Accomplishments that we're proud of

I'm particularly proud of creating a clean, intuitive user interface that makes complex gaming statistics accessible and engaging to view. Successfully implementing AI-powered analysis that provides meaningful, personalized insights rather than generic feedback was a major achievement. The responsive design works seamlessly across devices, and the visual categorization of different stat types (core stats, fun stats, performance stats) creates an excellent user experience that rivals commercial gaming analytics platforms.

What we learned

This project taught me valuable lessons about working with external APIs, particularly around data structure validation and error handling when dealing with inconsistent data formats. I gained significant experience with AWS serverless architecture and learned how to effectively integrate multiple AWS services like Lambda and Bedrock. Working with AI models for structured text generation showed me the importance of prompt engineering and how different models require different parameter formats and approaches.

What's next for Blitz Review

Future enhancements include expanding the analysis to cover multiple seasons and implementing trend tracking over time to show player improvement or decline. I plan to add more detailed champion-specific insights, including matchup analysis and build recommendations based on performance data. Integration with additional gaming platforms beyond League of Legends would make this a comprehensive gaming analytics tool, and adding social features like sharing summaries or comparing stats with friends would increase user engagement.

Built With

• amazon-web-services
• amplify
• bedrock
• lambda
• python
• riotapi
• s3