fehlende_werte.png

# ===========================================================================
# MODULE 06 · Missing values
# ===========================================================================
print("[06] Fehlende Werte ...")

# All relevant columns (cohort + labs, excluding patient_id)
all_cols = df.drop(columns=["patient_id"])
missing_pct = all_cols.isna().mean().mul(100).sort_values(ascending=True)
missing_pct = missing_pct[missing_pct > 0]  # only columns with missing values

fig, ax = plt.subplots(figsize=(8, max(3.5, len(missing_pct) * 0.55)))

bar_colors = [EVENT if pct >= 15 else PRIMARY if pct >= 5 else SECONDARY
              for pct in missing_pct.values]
bars = ax.barh(missing_pct.index, missing_pct.values, color=bar_colors,
               edgecolor="none", height=0.6)
ax.set_xlabel("Anteil fehlender Werte (%)")
ax.set_title("Fehlende Werte je Spalte")
ax.grid(axis="x")
ax.grid(axis="y", visible=False)
ax.set_xlim(0, missing_pct.max() * 1.30)

for bar, pct in zip(bars, missing_pct.values):
    ax.text(pct + 0.3, bar.get_y() + bar.get_height() / 2,
            f"{pct:.1f} %", va="center", fontsize=9.5)

# Legend
from matplotlib.patches import Patch
legend_items = [
    Patch(facecolor=EVENT,     label="≥15 % (klinisch informativ)"),
    Patch(facecolor=PRIMARY,   label="5–15 %"),
    Patch(facecolor=SECONDARY, label="< 5 %"),
]
ax.legend(handles=legend_items, loc="lower right", fontsize=9)

save(fig, ASSETS / "06-transformation" / "assets" / "fehlende_werte.png")