import matplotlib.pyplot as plt
import pandas as pd
import numpy as np
import seaborn as sns

# Data
data = {
    "Category": [
        "Business (all entries)", "Construction", "Entertainment", "Finance", "Healthcare",
        "Hospitality", "Legal", "Manufacturer", "Marketing", "Media", "Other",
        "Product", "Real estate", "Resource extraction", "Retail", "Technology", "Transportation"
    ],
    "Total": [142, 2, 7, 9, 7, 11, 1, 12, 1, 5, 3, 13, 3, 4, 12, 47, 5],
    "Approved": [22, 0, 1, 1, 3, 3, 0, 1, 0, 1, 0, 3, 1, 2, 1, 3, 2],
    "Declined": [117, 2, 6, 8, 4, 7, 1, 11, 1, 4, 3, 10, 2, 2, 11, 42, 3],
    "Draft": [2, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 2, 0],
    "Deleted/Other": [1, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0],
    "Approval%": [15, 0, 14, 11, 43, 27, 0, 8, 0, 20, 0, 23, 33, 50, 8, 6, 40],
    "Decline%": [82, 100, 86, 89, 57, 64, 100, 92, 100, 80, 100, 77, 67, 50, 92, 89, 60],
}
df = pd.DataFrame(data)

sns.set(style="whitegrid")
palette = sns.color_palette(["#4CAF50", "#F44336", "#FFC107", "#9E9E9E"])

fig = plt.figure(constrained_layout=False, figsize=(12, 10))
gs = fig.add_gridspec(3, 2, height_ratios=[1, 1, 0.7], width_ratios=[1, 0.02], hspace=0.35)

# 1) Stacked bar: Approved / Declined / Others
ax0 = fig.add_subplot(gs[0, 0])
ind = np.arange(len(df))
width = 0.6
p1 = ax0.bar(ind, df["Approved"], width, label="Approved", color=palette[0])
p2 = ax0.bar(ind, df["Declined"], width, bottom=df["Approved"], label="Declined", color=palette[1])
bottom2 = df["Approved"] + df["Declined"]
p3 = ax0.bar(ind, df["Draft"], width, bottom=bottom2, label="Draft", color=palette[2])
p4 = ax0.bar(ind, df["Deleted/Other"], width, bottom=bottom2 + df["Draft"], label="Deleted/Other", color=palette[3])

ax0.set_xticks(ind)
ax0.set_xticklabels(df["Category"], rotation=45, ha="right", fontsize=9)
ax0.set_ylabel("Count")
ax0.set_title("Submission Outcomes by Business Category (stacked)")
ax0.legend(frameon=False, fontsize=9)
ax0.set_ylim(0, df["Total"].max() * 1.08)
for i, tot in enumerate(df["Total"]):
    ax0.text(i, tot + df["Total"].max() * 0.01, str(tot), ha="center", fontsize=8, weight="bold")

# 2) Approval vs Decline % grouped bars
ax1 = fig.add_subplot(gs[1, 0])
w = 0.35
ax1.bar(ind - w/2, df["Approval%"], w, label="Approval %", color="#1976D2")
ax1.bar(ind + w/2, df["Decline%"], w, label="Decline %", color="#D32F2F")
ax1.set_xticks(ind)
ax1.set_xticklabels(df["Category"], rotation=45, ha="right", fontsize=9)
ax1.set_ylabel("Percentage (%)")
ax1.set_ylim(0, 105)
ax1.set_title("Approval vs Decline Percentage by Business Category")
ax1.legend(frameon=False, fontsize=9)
for i, (a, d) in enumerate(zip(df["Approval%"], df["Decline%"])):
    ax1.text(i - w/2, a + 3, f"{a}%", ha="center", fontsize=8)
    ax1.text(i + w/2, d + 3, f"{d}%", ha="center", fontsize=8)

# 3) Horizontal bar: Declined count proportion (sorted) to highlight problem areas
ax2 = fig.add_subplot(gs[2, 0])
decline_prop = df["Declined"] / df["Total"] * 100
order = decline_prop.sort_values(ascending=False).index
y_pos = np.arange(len(df))
ax2.barh(y_pos, decline_prop.iloc[order], color="#F44336")
ax2.set_yticks(y_pos)
ax2.set_yticklabels(df["Category"].iloc[order], fontsize=9)
ax2.set_xlabel("Declined as % of Total")
ax2.set_title("Decline Proportion by Business Category (sorted)")
for i, val in enumerate(decline_prop.iloc[order]):
    ax2.text(val + 1, i, f"{val:.0f}%", va="center", fontsize=8)

plt.tight_layout(rect=[0, 0, 1, 0.98])
plt.show()