1 of 10

TEAM OSPRID · AQX SPORTS ANALYTICS HACKATHON 2026

COURT

IQ

NBA ANALYTICS PLATFORM

Real-time scores · Shot quality models · Advanced metrics

Player comparisons · What-If simulator

🏀

BUILT FOR

Coaches

Front Offices

Fan Analysts

Ashok Pasala

Full Stack Develper

Snigdha Gorai

AI / ML Engineer

github.com/teamosprid/courtiq · Deployed on Vercel + Render

2 of 10

THE PROBLEM

NBA data exists.

No one makes it accessible.

📊

Fragmented Data

Stats are scattered across NBA.com, Basketball-Reference, ESPN — no unified view.

🧠

Surface-Level Metrics

PPG and FG% dominate coverage. Advanced models like xPTS are locked behind paywalls.

No Shot Intelligence

Coaches and fans can't instantly see which shots are smart vs. costly — without a data team.

CourtIQ solves all three — in a single, free, open-source web platform.

3 of 10

THE SOLUTION

CourtIQ — A Unified NBA Platform

📺

Live Scoreboard

Today's games, quarter scores, team leaders — auto-refreshing.

👤

Player Profiles

7-tab deep dive: stats, shot chart, game log, advanced metrics, hustle, splits, career.

🏆

Standings

Real-time East/West with W/L/PCT/GB/streak and playoff picture.

📈

League Leaders

8-stat leaderboards — PTS, REB, AST, STL, BLK, TS%, 3PM, PER.

🎯

Shot Quality (xPTS)

ML model scoring every shot on location, defender distance, shot clock, and dribbles.

What-If Simulator

Drag sliders to change a player's shot mix — see projected PPG update live.

📊

Player Compare

Head-to-head on 15 stats with radar chart and edge summary.

4 of 10

CORE INNOVATION

The xPTS

Shot Quality Model

Every shot is graded — not just made or missed.

xPTS = Expected FG% × Point Value of Shot

Expected FG% trained on:

Shot zone & distance · Defender distance · Shot clock remaining · Dribbles before shot · Shot type

MODEL INPUTS

📍

Shot Zone

🛡️

Defender Dist.

⏱️

Shot Clock

🏀

Dribbles

🎯

Shot Type

XGBoost Classifier · Trained on nba_api shot logs · Served via FastAPI

SHOT ZONE BREAKDOWN — STEPHEN CURRY 2024-25

Zone

FGA

Pts/Shot

Rating

Restricted Area

312

1.36

ELITE

Paint (Non-RA)

98

0.84

AVG

Mid-Range

142

0.89

POOR

Left Corner 3

88

1.30

ELITE

Right Corner 3

91

1.35

ELITE

Above Break 3

386

1.23

GOOD

Left Baseline

42

0.76

POOR

5 of 10

ARCHITECTURE

Full-Stack Technical Design

DATA LAYER

nba_api

Shot logs · Live scores · Player tracking · Play-by-play · Hustle stats

Basketball-Ref

Advanced metrics · Career stats · Season averages · Win shares

NBA Tracking

Defender distance · Touch time · Dribbles · Miles covered · Speed

BACKEND

FastAPI (Python)

REST endpoints · Async · Auto-docs · 8 routes serving all data

XGBoost Model

xPTS classifier · Trained on 200K+ shots · Shot quality scoring

Data Pipeline

Fetch → clean → cache (CSV/parquet) → serve · Runs on Render (free)

FRONTEND

React 18 + Vite

7 pages · React Router · Component architecture · Fast HMR

Tailwind CSS

Custom NBA black/accent theme · Responsive · Zero runtime overhead

Recharts + SVG

Line charts · Radar charts · Interactive court heatmap · Animations

Vercel · React Frontend

Free · Auto-deploy from GitHub · CDN

Render · FastAPI Backend

Free · Docker · Auto-sleep on idle

GitHub · Open Source Repo

Public · MIT License · Full source

6 of 10

DATA DEPTH

Every Metric That Matters

200+

METRICS TRACKED

30

NBA TEAMS

450+

ACTIVE PLAYERS

82

GAMES / SEASON

7

DASHBOARD PAGES

Live Game Data

Real-time scores & quarter breakdown

Live play-by-play feed

Shot tracking per game

Team leaders per quarter

Broadcast & arena info

Player Analytics

Shot coordinates (x,y on court)

Defender distance & shot clock

Hustle: deflections, charges, miles

On/Off splits & matchup data

Clutch performance stats

Advanced Metrics

PER · Win Shares · VORP · BPM

Offensive & Defensive Rating

True Shooting % · Usage Rate

Net Rating & Four Factors

Season splits & trend lines

xPTS Model Output

Expected FG% per shot

xPTS — expected points per shot

Actual vs expected overperformance

Zone ratings: ELITE/GOOD/AVG/POOR

What-If projected PPG simulator

7 of 10

FEATURE SPOTLIGHT

The What-If Shot Simulator

No other free platform lets coaches simulate shot-mix changes and see the projected scoring impact in real time.

How It Works

01

Select a player

Choose any NBA player from the dropdown.

02

Adjust shot distribution sliders

Drag to change % of shots from Rim / Corner 3 / Above Break 3 / Mid-Range.

03

Model recalculates instantly

xPTS model applies zone-specific efficiency to new shot mix.

04

See projected PPG update live

Dashboard shows ▲ or ▼ change vs current season baseline.

EXAMPLE OUTPUT — STEPHEN CURRY

At Rim %

35%

Corner 3 %

22%

Above Break 3 %

35%

Mid-Range %

8%

PROJECTED PPG

28.7

▲ +2.3 PPG from optimizing shot mix

8 of 10

WHY WE WIN

Judging Criteria — How We Score

01 Analytical Insight

Highest weight

— xPTS model: statistically sound, uses 5 features per shot

— Zone breakdown: every zone rated vs league expectations

— Expected FG% vs actual — reveals true shooting skill

— Monthly trend charts show shot quality evolution

— Backed by 200K+ real NBA shots from nba_api

02 Practical Application

High weight

— Coaches: instantly see which shots a player should stop taking

— Front offices: compare player shot profiles before trades

— What-If Simulator: concrete, actionable PPG projections

— Injury report integrated for lineup decisions

— Hustle stats (deflections, charges, miles) for scouting

03 Data Presentation

High weight

— Live interactive web app — no slides, no static charts

— Color-coded court heatmap: instantly readable by anyone

— ELITE/GOOD/AVG/POOR zone ratings at a glance

— Live ticker, quarter breakdown, real-time score bars

— Mobile-responsive, professional NBA-grade design

9 of 10

THE TEAM

Team

Osprid

AQX Sports Analytics Hackathon

2026 — Open Source Division

TECH USED

· React 18 · Vite · Tailwind CSS

· FastAPI · Python · XGBoost

· nba_api · Data Engineering

· SVG Visualization · Recharts

· Vercel · Render · GitHub Actions

AP

Ashok Pasala

Full Stack Developer

React · Vite · Tailwind · FastAPI

API design · UI architecture

Deployment (Vercel + Render)

github.com/ashokpasala

SG

Snigdha Gorai

AI / ML Engineer

XGBoost · scikit-learn · Python

xPTS model design & training

Data pipeline (nba_api · pandas)

github.com/snigdhagorai

10 of 10

🏀

COURT

IQ

NBA ANALYTICS PLATFORM

by Team Osprid

THANK YOU

Ready to

Change the Game.

✓ Live deployed web app — open right now

✓ Custom ML model (xPTS) — real NBA data

✓ 7 pages of analytics in one platform

✓ What-If simulator — unique feature

✓ 100% open source — full GitHub repo

✓ Built in 3 days by 2 people

🔗 https://github.com/Snigdha-0210/Courtiq

🌐 https://courtiq-eight-gamma.vercel.app/