KRAMER - Kanaria RArity MEteR

Inspiration

Many rarity services are represented on the Internet. We bring the best of our experience, knowledge, and intuition to design interpretable rarity meter.

What it does

KRAMER is a rarity meter for the Kanaria collection. KRAMER takes birds' individual traits, their sets, and edition and fits their combination to trade data. We've written an article on Medium about the score computation and its intuition.

With the KRAMER you can

get the score and the rank of the wanted bird at https://www.kramer.tools, for example, https://www.kramer.tools/0436
have a look at the whole collection at https://www.kramer.tools//allStats.

How we built it

The source code is available in our repository. Follow the instructions from the Readme to build the project locally.

Under the hood

The data mining and machine learning parts are performed in IPython and notebooks are in https://github.com/yashmadhwal/Kanaria/tree/main/IPython. Namely,

ReadDataset: reads the Kanaria dataset and saves it as a bird-trait table
ReadTradeDataset: reads the trade data of the Kanaria dataset and saves it as an event list
ComputeIndex: computes scores based on traits, sets, and edition
FitWeights: fits the best combination of traits, sets, and edition scores to trade data
GetJPEGCollections: downloads all the birds' images.

The web app part is performed on Python (Flask, iPython) and Nodejs. Please, follow the instructions from Readme to build it.