app.py

import joblib
import os
import streamlit as st
import pandas as pd
import numpy as np
from sklearn.model_selection import train_test_split
from sklearn.preprocessing import LabelEncoder
from sklearn.ensemble import RandomForestClassifier
from sklearn.metrics import accuracy_score, confusion_matrix
import matplotlib.pyplot as plt

st.title("👨‍💼 Employee Performance Predictor")

# Upload file
uploaded_file = st.file_uploader("Upload Employee Dataset", type=["csv"])

if uploaded_file is not None:

    df = pd.read_csv(uploaded_file)

    st.subheader("📊 Dataset Preview")
    st.write(df.head())

    # ---------------------------
    # Encoding
    # ---------------------------
    dept_encoder = LabelEncoder()
    perf_encoder = LabelEncoder()

    df['Department'] = dept_encoder.fit_transform(df['Department'])
    df['Performance'] = perf_encoder.fit_transform(df['Performance'])

    # ---------------------------
    # Features & Target
    # ---------------------------
    X = df.drop('Performance', axis=1)
    y = df['Performance']

    # Train-test split
    X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.2)

    # ---------------------------
    # Folders
    # ---------------------------
    os.makedirs("models", exist_ok=True)
    os.makedirs("images", exist_ok=True)

    # ---------------------------
    # Model (Load or Train)
    # ---------------------------
    model_path = "models/performance_model.pkl"

    if os.path.exists(model_path):
        model = joblib.load(model_path)
    else:
        model = RandomForestClassifier()
        model.fit(X_train, y_train)
        joblib.dump(model, model_path)

    # ---------------------------
    # Prediction
    # ---------------------------
    y_pred = model.predict(X_test)

    # Accuracy
    acc = accuracy_score(y_test, y_pred)

    st.subheader("📈 Model Accuracy")
    st.write(f"Accuracy: {acc:.2f}")

    # =====================================================
    # 📉 CONFUSION MATRIX (FIXED + PROPER SIZE)
    # =====================================================
    cm = confusion_matrix(y_test, y_pred)
    st.subheader("📉 Confusion Matrix")

    fig, ax = plt.subplots(figsize=(7, 7))  # ✅ bigger size

    im = ax.imshow(cm, cmap="Blues")

    # Labels
    ax.set_xlabel("Predicted")
    ax.set_ylabel("Actual")

    classes = perf_encoder.classes_
    ax.set_xticks(range(len(classes)))
    ax.set_yticks(range(len(classes)))
    ax.set_xticklabels(classes, rotation=45)
    ax.set_yticklabels(classes)

    # Values inside matrix
    for i in range(len(classes)):
        for j in range(len(classes)):
            ax.text(j, i, cm[i, j],
                    ha="center", va="center",
                    color="black", fontsize=12)

    fig.colorbar(im)

    # Save image
    fig.savefig("images/confusion_matrix.png", bbox_inches="tight", dpi=300)

    st.pyplot(fig)

    # =====================================================
    # 🔥 FEATURE IMPORTANCE
    # =====================================================
    st.subheader("🔥 Feature Importance")

    importance = model.feature_importances_
    features = X.columns

    fig2, ax2 = plt.subplots(figsize=(7, 4))
    ax2.barh(features, importance)
    ax2.set_xlabel("Importance")

    fig2.savefig("images/feature_importance.png", bbox_inches="tight", dpi=300)

    st.pyplot(fig2)

    # =====================================================
    # 🔮 PREDICTION
    # =====================================================
    st.subheader("🔮 Predict New Employee Performance")

    age = st.slider("Age", 20, 60)
    exp = st.slider("Experience", 1, 20)
    salary = st.number_input("Salary", min_value=10000)
    training = st.slider("Training Hours", 1, 100)

    dept = st.selectbox("Department", list(dept_encoder.classes_))
    dept_encoded = dept_encoder.transform([dept])[0]

    if st.button("Predict"):

        input_df = pd.DataFrame(
            [[age, exp, salary, training, dept_encoded]],
            columns=X.columns
        )

        pred = model.predict(input_df)

        result = perf_encoder.inverse_transform(pred)

        st.success(f"Predicted Performance: {result[0]}")