MarkusNeusinger
diff --git a/‎plots/line-win-probability/implementations/plotnine.py‎
Lines changed: 177 additions & 0 deletions b/‎plots/line-win-probability/implementations/plotnine.py‎
Lines changed: 177 additions & 0 deletions
@@ -0,0 +1,177 @@
+""" pyplots.ai
+line-win-probability: Win Probability Chart
+Library: plotnine 0.15.3 | Python 3.14.3
+Quality: 92/100 | Created: 2026-03-20
+"""
+
+import numpy as np
+import pandas as pd
+from plotnine import (
+    aes,
+    annotate,
+    coord_cartesian,
+    element_blank,
+    element_line,
+    element_rect,
+    element_text,
+    geom_hline,
+    geom_line,
+    geom_point,
+    geom_rect,
+    geom_ribbon,
+    geom_segment,
+    geom_text,
+    ggplot,
+    labs,
+    scale_alpha_identity,
+    scale_fill_manual,
+    scale_x_continuous,
+    scale_y_continuous,
+    theme,
+    theme_minimal,
+)
+
+
+# Data
+np.random.seed(42)
+
+n_plays = 130
+plays = np.arange(n_plays)
+win_prob = np.zeros(n_plays)
+win_prob[0] = 0.50
+
+scoring_plays = {
+    12: ("FG Home", 0.10),
+    28: ("TD Away", -0.18),
+    42: ("TD Home", 0.22),
+    55: ("FG Away", -0.08),
+    68: ("TD Home", 0.15),
+    82: ("TD Away", -0.20),
+    95: ("FG Home", 0.12),
+    110: ("TD Home", 0.16),
+    122: ("FG Away", -0.05),
+}
+
+events = {}
+for i in range(1, n_plays):
+    drift = np.random.normal(0, 0.012)
+    if i in scoring_plays:
+        label, shift = scoring_plays[i]
+        win_prob[i] = win_prob[i - 1] + shift + drift
+        events[i] = label
+    else:
+        win_prob[i] = win_prob[i - 1] + drift
+
+win_prob = np.clip(win_prob, 0.04, 0.96)
+
+for i in range(n_plays - 8, n_plays):
+    t = (i - (n_plays - 8)) / 7.0
+    win_prob[i] = win_prob[n_plays - 9] * (1 - t) + 0.78 * t
+
+home_fill = np.maximum(win_prob, 0.5)
+away_fill = np.minimum(win_prob, 0.5)
+
+df = pd.DataFrame({"play": plays, "win_prob": win_prob})
+
+df_home = pd.DataFrame({"play": plays, "ymin": 0.5, "ymax": home_fill, "team": "Eagles (Home)"})
+df_away = pd.DataFrame({"play": plays, "ymin": away_fill, "ymax": 0.5, "team": "Cowboys (Away)"})
+df_ribbon = pd.concat([df_home, df_away], ignore_index=True)
+
+event_df = pd.DataFrame(
+    {"play": list(events.keys()), "win_prob": [win_prob[p] for p in events.keys()], "label": list(events.values())}
+)
+
+# Smart annotation positioning with staggered offsets to avoid overlap in Q4
+label_offsets = []
+for _idx, row in event_df.iterrows():
+    base = 0.06 if row["win_prob"] > 0.5 else -0.06
+    # Extra offset for crowded late-game region
+    if row["play"] >= 90:
+        base *= 1.5
+    label_offsets.append(row["win_prob"] + base)
+event_df["label_y"] = label_offsets
+
+# Highlight the decisive moment (TD Home at play 110 that sealed the game)
+decisive_play = 110
+highlight_df = pd.DataFrame({"xmin": [104], "xmax": [116], "ymin": [0.50], "ymax": [0.96], "alpha": [0.06]})
+
+# Quarter boundary data for geom_segment (plotnine-idiomatic layer composition)
+quarter_df = pd.DataFrame({"x": [32, 65, 97], "ymin": [0.0] * 3, "ymax": [1.0] * 3})
+
+# Plot
+quarter_breaks = [0, 32, 65, 97, 129]
+quarter_labels = ["Kickoff", "Q2", "Halftime", "Q4", "Final"]
+
+plot = (
+    ggplot()
+    # Decisive moment highlight zone
+    + geom_rect(
+        aes(xmin="xmin", xmax="xmax", ymin="ymin", ymax="ymax", alpha="alpha"),
+        data=highlight_df,
+        fill="#DAA520",
+        inherit_aes=False,
+    )
+    + scale_alpha_identity()
+    # Team-colored area fills
+    + geom_ribbon(aes(x="play", ymin="ymin", ymax="ymax", fill="team"), data=df_ribbon, alpha=0.35)
+    # 50% reference line
+    + geom_hline(yintercept=0.5, color="#888888", size=0.6, linetype="dashed")
+    # Quarter boundaries via geom_segment (plotnine-idiomatic vs geom_vline)
+    + geom_segment(
+        aes(x="x", xend="x", y="ymin", yend="ymax"),
+        data=quarter_df,
+        color="#cccccc",
+        size=0.4,
+        linetype="dotted",
+        inherit_aes=False,
+    )
+    # Win probability trace
+    + geom_line(aes(x="play", y="win_prob"), data=df, color="#1a1a1a", size=1.2)
+    # Scoring event markers
+    + geom_point(
+        aes(x="play", y="win_prob"), data=event_df, color="#1a1a1a", size=4, fill="white", stroke=0.8, shape="o"
+    )
+    # Event annotations
+    + geom_text(aes(x="play", y="label_y", label="label"), data=event_df, size=7, fontweight="bold", color="#333333")
+    # Scales
+    + scale_fill_manual(values={"Eagles (Home)": "#004C54", "Cowboys (Away)": "#8B1A1A"})
+    + scale_x_continuous(breaks=quarter_breaks, labels=quarter_labels, expand=(0.03, 2))
+    + scale_y_continuous(
+        labels=lambda lst: [f"{int(v * 100)}%" for v in lst], limits=(0, 1), breaks=[0, 0.25, 0.5, 0.75, 1.0]
+    )
+    + coord_cartesian(xlim=(-2, 134))
+    + labs(
+        x="Game Progression", y="Home Win Probability", title="line-win-probability · plotnine · pyplots.ai", fill=""
+    )
+    # Final score box - placed bottom-left for balance
+    + annotate(
+        "label",
+        x=10,
+        y=0.06,
+        label="Final: Eagles 24 – Cowboys 17",
+        size=9,
+        fill="#f5f5f5",
+        color="#333333",
+        fontweight="bold",
+        label_padding=0.5,
+    )
+    # Theme with plotnine-specific element styling
+    + theme_minimal()
+    + theme(
+        figure_size=(16, 9),
+        plot_title=element_text(size=24, weight="bold"),
+        axis_title_x=element_text(size=18),
+        axis_title_y=element_text(size=18),
+        axis_text=element_text(size=16, color="#555555"),
+        legend_text=element_text(size=15),
+        legend_position="top",
+        legend_background=element_rect(fill="#fafafa", color="#e0e0e0", size=0.3),
+        legend_key_size=18,
+        panel_grid_major_x=element_blank(),
+        panel_grid_minor=element_blank(),
+        panel_grid_major_y=element_line(color="#e8e8e8", size=0.3),
+    )
+)
+
+# Save
+plot.save("plot.png", dpi=300, verbose=False)