app.py

import streamlit as st
import pandas as pd
import matplotlib.pyplot as plt
import matplotlib.ticker as mticker
import seaborn as sns

# ── Page config ──────────────────────────────────────────────────────────────
st.set_page_config(
    page_title="Superstore Sales Analytics",
    page_icon="📊",
    layout="wide",
)

# ── Global style ─────────────────────────────────────────────────────────────
PALETTE = ["#4C9BE8", "#E8834C", "#4CE882", "#E84C6B", "#A04CE8", "#E8D14C"]
sns.set_theme(style="darkgrid", palette=PALETTE)
plt.rcParams.update({
    "figure.facecolor": "#0e1117",
    "axes.facecolor":   "#262730",
    "axes.labelcolor":  "#ccc",
    "xtick.color":      "#ccc",
    "ytick.color":      "#ccc",
    "text.color":       "#eee",
    "grid.color":       "#3a3a4a",
})

# ── Load data ─────────────────────────────────────────────────────────────────
@st.cache_data
def load_data():
    for enc in ("windows-1252", "latin-1", "utf-8"):
        try:
            df = pd.read_csv("Superstore.csv", encoding=enc)
            break
        except UnicodeDecodeError:
            continue
    df["Order Date"] = pd.to_datetime(df["Order Date"], errors="coerce")
    df["Ship Date"]  = pd.to_datetime(df["Ship Date"],  errors="coerce")
    df["Year"]        = df["Order Date"].dt.year
    df["Month"]       = df["Order Date"].dt.month
    df["YearMonth"]   = df["Order Date"].dt.to_period("M").astype(str)
    df["Profit Margin %"] = (df["Profit"] / df["Sales"].replace(0, pd.NA) * 100).round(2)
    return df

df = load_data()

# ── Sidebar Filters ───────────────────────────────────────────────────────────
st.sidebar.image("https://img.icons8.com/color/96/combo-chart--v2.png", width=60)
st.sidebar.title("Filters")

years = sorted(df["Year"].dropna().unique().astype(int))
sel_years = st.sidebar.multiselect("Year", years, default=years)

regions = sorted(df["Region"].dropna().unique())
sel_regions = st.sidebar.multiselect("Region", regions, default=regions)

categories = sorted(df["Category"].dropna().unique())
sel_cats = st.sidebar.multiselect("Category", categories, default=categories)

# Apply filters
mask = (
    df["Year"].isin(sel_years) &
    df["Region"].isin(sel_regions) &
    df["Category"].isin(sel_cats)
)
fdf = df[mask]

st.sidebar.divider()
st.sidebar.caption(f"Showing **{len(fdf):,}** of **{len(df):,}** rows")

# ── Header ────────────────────────────────────────────────────────────────────
st.title("📊 Superstore Sales Analytics")
st.markdown("An interactive end-to-end analytics dashboard built with Python & Streamlit.")
st.divider()

# ── KPI Row ───────────────────────────────────────────────────────────────────
total_sales   = fdf["Sales"].sum()
total_profit  = fdf["Profit"].sum()
total_orders  = fdf["Order ID"].nunique()
avg_order_val = total_sales / total_orders if total_orders else 0
profit_margin = (total_profit / total_sales * 100) if total_sales else 0

k1, k2, k3, k4, k5 = st.columns(5)
k1.metric("💰 Total Sales",        f"${total_sales:,.0f}")
k2.metric("📈 Total Profit",       f"${total_profit:,.0f}")
k3.metric("🛒 Total Orders",       f"{total_orders:,}")
k4.metric("🧾 Avg. Order Value",   f"${avg_order_val:,.0f}")
k5.metric("📊 Profit Margin",      f"{profit_margin:.1f}%")
st.divider()

# ─── Helper: chart wrapper ────────────────────────────────────────────────────
def show_chart(fig):
    fig.tight_layout()
    st.pyplot(fig)
    plt.close(fig)

# ── Chart 1: Monthly Sales Trend ──────────────────────────────────────────────
st.subheader("📅 Monthly Sales Trend")
if not fdf.empty:
    trend = fdf.groupby("YearMonth")["Sales"].sum().reset_index()
    trend["Sales_K"] = trend["Sales"] / 1000

    fig, ax = plt.subplots(figsize=(14, 4))
    ax.fill_between(trend["YearMonth"], trend["Sales_K"], alpha=0.15, color=PALETTE[0])
    ax.plot(trend["YearMonth"], trend["Sales_K"], marker="o", ms=4,
            color=PALETTE[0], linewidth=2)
    ax.set_xlabel("Month")
    ax.set_ylabel("Sales ($ thousands)")
    ax.yaxis.set_major_formatter(mticker.FuncFormatter(lambda x, _: f"${x:.0f}K"))

    # Reduce x-tick clutter
    step = max(1, len(trend) // 12)
    ax.set_xticks(range(0, len(trend), step))
    ax.set_xticklabels(trend["YearMonth"].iloc[::step], rotation=45, ha="right", fontsize=8)
    show_chart(fig)
else:
    st.info("No data available for the selected filters.")

st.divider()

# ── Charts Row 2: Category & Region ──────────────────────────────────────────
col_a, col_b = st.columns(2)

with col_a:
    st.subheader("🗂️ Sales & Profit by Category")
    if not fdf.empty:
        cat = fdf.groupby("Category")[["Sales", "Profit"]].sum().reset_index()
        cat_melted = cat.melt(id_vars="Category", var_name="Metric", value_name="Amount")
        fig, ax = plt.subplots(figsize=(6, 4))
        sns.barplot(data=cat_melted, x="Category", y="Amount", hue="Metric",
                    ax=ax, palette=[PALETTE[0], PALETTE[2]])
        ax.set_xlabel("")
        ax.yaxis.set_major_formatter(mticker.FuncFormatter(lambda x, _: f"${x/1000:.0f}K"))
        ax.legend(facecolor="#262730", edgecolor="none")
        show_chart(fig)

with col_b:
    st.subheader("🌍 Sales by Region")
    if not fdf.empty:
        reg = fdf.groupby("Region")["Sales"].sum().reset_index().sort_values("Sales", ascending=False)
        fig, ax = plt.subplots(figsize=(6, 4))
        bars = sns.barplot(data=reg, x="Region", y="Sales", ax=ax,
                           hue="Region", legend=False, palette=PALETTE)
        ax.yaxis.set_major_formatter(mticker.FuncFormatter(lambda x, _: f"${x/1000:.0f}K"))
        ax.set_xlabel("")
        show_chart(fig)

st.divider()

# ── Charts Row 3: Sub-Category & Segment ──────────────────────────────────────
col_c, col_d = st.columns(2)

with col_c:
    st.subheader("📦 Profit by Sub-Category")
    if not fdf.empty:
        sub = (fdf.groupby("Sub-Category")["Profit"].sum()
                  .reset_index()
                  .sort_values("Profit", ascending=True))
        colors = [PALETTE[2] if v >= 0 else PALETTE[3] for v in sub["Profit"]]
        fig, ax = plt.subplots(figsize=(6, 6))
        ax.barh(sub["Sub-Category"], sub["Profit"], color=colors)
        ax.axvline(0, color="#ccc", linewidth=0.8)
        ax.xaxis.set_major_formatter(mticker.FuncFormatter(lambda x, _: f"${x/1000:.0f}K"))
        ax.set_xlabel("Profit ($)")
        show_chart(fig)

with col_d:
    st.subheader("👤 Sales Share by Customer Segment")
    if not fdf.empty:
        seg = fdf.groupby("Segment")["Sales"].sum()
        fig, ax = plt.subplots(figsize=(6, 6))
        wedges, texts, autotexts = ax.pie(
            seg.values,
            labels=seg.index,
            autopct="%1.1f%%",
            colors=PALETTE[:len(seg)],
            startangle=90,
            wedgeprops=dict(linewidth=2, edgecolor="#0e1117"),
        )
        for t in texts + autotexts:
            t.set_color("#eee")
        ax.set_facecolor("#0e1117")
        show_chart(fig)

st.divider()

# ── Raw Data Explorer ──────────────────────────────────────────────────────────
with st.expander("🔍 Explore Raw Data", expanded=False):
    cols_to_show = ["Order ID", "Order Date", "Customer Name", "Segment",
                    "Region", "Category", "Sub-Category", "Product Name",
                    "Sales", "Quantity", "Discount", "Profit", "Profit Margin %"]
    existing = [c for c in cols_to_show if c in fdf.columns]
    st.dataframe(
        fdf[existing].sort_values("Order Date", ascending=False).reset_index(drop=True),
        use_container_width=True,
        height=350,
    )
    csv = fdf[existing].to_csv(index=False).encode("utf-8")
    st.download_button("⬇️ Download Filtered Data as CSV", csv, "filtered_superstore.csv", "text/csv")

st.caption("Data: Superstore Sales Dataset  |  Built with Python, Pandas & Streamlit")