Criclysis

Cricket exploratory analysis and prediction platform. Let’s entertain audience interesting insights and match predictions.

Comment

Inspiration

We were trying to analyse cricket data using multiple websites to analyse predict best 11, winners and great performances but at we used to lose insights because we were moving around many sources and websites.

What it does

Currently it provides aggregated data at multiple levels. Player level, match level, tournament level and whole cricket level. In future we will expand it to predict matches, best 11 and performances.

How we built it

  • We collected data from cricsheet.org in json format
  • Processed data using pyspark at scale
  • pushed data AWS S3 storage
  • created Aurora PostgreSql database
  • Created tables and inserted data present in S3
  • Gave a good prompt to Vercel V0 app to create the website
  • We were fascinated with the sophisticated and coherent website developed vercel agents
  • There were few issues with initial versions
  • Iterated development using well crafted prompts create the final version of website.

Challenges we ran into

  • Processing and automating data insert into S3
  • Building tables in Aurora using data present in S3
  • Networking issue between front end and database
  • Crafting good prompts

Accomplishments that we're proud of

  • ultimate combination of next.js and Aurora PostgreSql
  • Exploratory analysis done during development
  • Providing useful stats to end users
  • Vibe coding and secure design principles

What we learned

  • Deploying AWS and vercel solution
  • Crafting good prompts
  • Vibe coding
  • Networking to bring data between different platforms
  • Vercel skills
  • next.js, typescript
  • AWS
  • Time management and goal setting

What's next for Criclysis

  • Building Prediction engine
  • AI assisted bot
  • Embedded RAG application
  • Full scale deployment with own domain
Share this project:

Updates