Tire Strategy Analysis

import tif1
import pandas as pd

session = tif1.get_session(2025, "British Grand Prix", "Race")
laps = session.laps

# Group by driver and stint to see strategy
strategy = laps.groupby(["Driver", "Stint", "Compound"]).agg({
    "LapNumber": ["min", "max", "count"]
}).reset_index()

print(strategy.head())

import matplotlib.pyplot as plt

plt.figure(figsize=(12, 8))
# session.drivers returns driver numbers; translate them to abbreviations
drivers = [session.get_driver(i)["Abbreviation"] for i in session.drivers[:10]]  # Top 10 drivers

for i, driver in enumerate(drivers):
    driver_laps = laps[laps["Driver"] == driver]

    for stint in driver_laps["Stint"].unique():
        stint_laps = driver_laps[driver_laps["Stint"] == stint]
        compound = stint_laps["Compound"].iloc[0]

        # Color mapping
        color = {"SOFT": "red", "MEDIUM": "yellow", "HARD": "white"}.get(compound, "cyan")

        plt.barh(y=driver,
                 width=stint_laps["LapNumber"].max() - stint_laps["LapNumber"].min() + 1,
                 left=stint_laps["LapNumber"].min(),
                 color=color,
                 edgecolor="black")

plt.title("Race Strategy - Top 10 Drivers")
plt.xlabel("Lap Number")
plt.ylabel("Driver")
plt.gca().invert_yaxis()
plt.show()

# Filter for a long stint (e.g., VER Stint 2)
ver_stint_2 = laps[(laps["Driver"] == "VER") & (laps["Stint"] == 2)]

# Simple linear regression to find degradation rate
# scipy needs numeric values, so convert the timedelta lap times to seconds
from scipy import stats
lap_seconds = ver_stint_2["LapTime"].dt.total_seconds()
slope, intercept, r_value, p_value, std_err = stats.linregress(
    ver_stint_2["TyreLife"],
    lap_seconds
)

print(f"Degradation Rate: {slope:.4f} seconds per lap")

# Average lap time per compound (cleaned)
clean_laps = laps[laps["LapTime"] < laps["LapTime"].min() * 1.1]
compound_performance = clean_laps.groupby("Compound")["LapTime"].mean()

print(compound_performance)