Delete pages 5_💲_Magenment_deck.py

pages/5_💲_Magenment_deck.py

@@ -1,73 +0,0 @@
-import streamlit as st
-import pandas as pd
-import numpy as np
-import joblib
-
-st.set_page_config(page_title="💲 Management Deck: Cost Analysis", layout="wide")
-st.title("💲 Management Deck: Cost Analysis of Model Decisions")
-
-# --- Load model and test data ---
-@st.cache_data
-def load_model_and_data():
-    model = joblib.load("model_1mvp.pkl")
-    df_test = pd.read_csv("test_data.csv")
-    return model, df_test
-
-model, df_test = load_model_and_data()
-target = "y"
-X_test = df_test.drop(columns=[target])
-y_test = df_test[target]
-
-# --- Sidebar: Cost selection ---
-with st.sidebar.form("cost_selector"):
-    st.markdown("## Set Cost Parameters")
-    cost_fp = st.number_input("Cost of False Positive (FP)", min_value=0, value=5, step=1)
-    cost_fn = st.number_input("Cost of False Negative (FN)", min_value=0, value=30, step=1)
-    submitted = st.form_submit_button("Update Cost Analysis")
-
-if submitted:
-    st.subheader("Cost Analysis Based on Model Predictions")
-
-    # Predict probabilities and classes
-    y_pred_proba = model.predict_proba(X_test)[:, 1]
-    threshold = 0.5
-    y_pred = (y_pred_proba >= threshold).astype(int)
-
-    # Confusion matrix components
-    FP = np.sum((y_pred == 1) & (y_test == 0))
-    FN = np.sum((y_pred == 0) & (y_test == 1))
-    TP = np.sum((y_pred == 1) & (y_test == 1))
-    TN = np.sum((y_pred == 0) & (y_test == 0))
-
-    total_cost = FP * cost_fp + FN * cost_fn
-
-    st.markdown(f"""
-    **Threshold:** {threshold:.2f}  
-    - **False Positives (FP):** {FP} × {cost_fp} = {FP * cost_fp}
-    - **False Negatives (FN):** {FN} × {cost_fn} = {FN * cost_fn}
-    - **True Positives (TP):** {TP}
-    - **True Negatives (TN):** {TN}
-    ---
-    ## **Total Cost: {total_cost}**
-    """)
-
-    # Optional: Show cost as a function of threshold
-    st.subheader("Cost vs. Threshold")
-    thresholds = np.linspace(0, 1, 120)
-    costs = []
-    for t in thresholds:
-        y_pred_t = (y_pred_proba >= t).astype(int)
-        FP_t = np.sum((y_pred_t == 1) & (y_test == 0))
-        FN_t = np.sum((y_pred_t == 0) & (y_test == 1))
-        costs.append(FP_t * cost_fp + FN_t * cost_fn)
-    import matplotlib.pyplot as plt
-    fig, ax = plt.subplots(figsize=(8, 4))
-    ax.plot(thresholds, costs, label="Total Cost")
-    ax.axvline(threshold, color="red", linestyle="--", label=f"Current threshold = {threshold:.2f}")
-    ax.set_xlabel("Threshold")
-    ax.set_ylabel("Total Cost")
-    ax.set_title("Total Cost vs. Classification Threshold")
-    ax.legend()
-    st.pyplot(fig)
-
-    st.caption("You can adjust the costs in the sidebar to see their impact on the total cost and optimal threshold.")