diamondfp

Comparing baseball players like molecules!

Statistical fingerprints for baseball players — find their closest matches.

diamondfp is a Python library for generating statistical fingerprints of baseball players, enabling similarity searches much like molecular similarity in chemistry. By encoding player performance into feature vectors, you can compare players across eras and styles using metrics like the Tanimoto coefficient/Jaccard Score.

Currently, diamondfp supports multiple types of fingerprint generation for batting, pitching, and fielding statistics.

Features

• Create statistical fingerprints from player performance data
• Compare players using similarity metrics (Tanimoto coefficient/Jaccard Score)
• Easily extendable to batting, pitching, and fielding statistics
• Works with historical and modern baseball data

Fingerprints

• Binary vector fingerprints
• Binned fingerprints
• Normalized continuous fingerprints (min-max & z-score normalization)

Usage

import pandas as pd
from diamondfp.fingerprints import binaryfp, binnedfp
from diamondfp.utils.features import generate_quantiles
from diamondfp.scoring import jaccard, manhattan, cosine_sim

df = pd.read_csv("data/career-batting.csv")

stat_features = {
    "H": [0.5, 0.75, 0.9, 0.95],
    "2B": [0.75, 0.95],
    "3B": [0.75, 0.95],
    "HR": [0.9, 0.99],
    "K%": [0.1, 0.25],
    "BB%": [0.75, 0.99],
    "AVG": [0.5, 0.75, 0.9, 0.95],
    "OBP": [0.5, 0.75, 0.9, 0.95],
    "SLG": [0.5, 0.75, 0.9, 0.95],
    "OPS": [0.5, 0.75, 0.9, 0.95],
}

feat_quants = generate_quantiles(df, stat_features)

babe_ruth = binaryfp(df[df["Name"] == "Babe Ruth"].squeeze(), feat_quants)
shohei_ohtani = binaryfp(df[df["Name"] == "Shohei Ohtani"].squeeze(), feat_quants)
sim_score = jaccard(babe_ruth, shohei_ohtani)
print(f"Jaccard score: {sim_score:0.2f}")  # 0.72
cos_sim = cosine_sim(babe_ruth, shohei_ohtani)
print(f"Cosine similarity: {cos_sim:0.2f}")  # 0.85
man_dist = manhattan(babe_ruth, shohei_ohtani)
print(f"Manhattan distance: {man_dist}")  # 5

Example Usage

Examples to how to use diamondfp can be found here!

Example notebooks use the Lahman Baseball Database, licensed under Creative Commons BY-SA 3.0.

Roadmap

• Binary vector fingerprints
• Binned fingerprints
• Clustering and visualization tools
• Clean up examples
• Code coverage
• PyPI release

License

This project is licensed under the MIT License. See the LICENSE file for details.