完成作業繳交 -楊舒媁- AIPE03

homework/m1_numpy.py

@@ -19,18 +19,24 @@
 def green_mean():
     """建立 [10, 20, 30, 40, 50]，回傳所有元素的平均值 (float)"""
     # TODO: 你的程式碼
+    float = np.array([10, 20, 30, 40, 50])
+    return np.mean(float)
     pass
 
 
 def green_double():
     """建立 [10, 20, 30, 40, 50]，回傳所有元素乘以 2 的 ndarray"""
     # TODO: 你的程式碼
+    ndarray = np.array([10, 20, 30, 40, 50])
+    return ndarray * 2
     pass
 
 
 def green_filter():
     """建立 [10, 20, 30, 40, 50]，回傳大於 25 的元素 (ndarray)"""
     # TODO: 你的程式碼
+    ndarray = np.array([10, 20, 30, 40, 50])
+    return ndarray[ndarray >25]
     pass
 
 
@@ -42,6 +48,7 @@ def green_filter():
 def yellow_expensive_count(prices):
     """回傳單價 > 1000 的商品數量 (int)"""
     # TODO: 你的程式碼
+    return (prices > 1000).sum()
     pass
 
 
@@ -51,6 +58,7 @@ def yellow_top3_stock_indices(stocks):
     提示：np.argsort
     """
     # TODO: 你的程式碼
+    return np.argsort(stocks)[::-1][:3]
     pass
 
 
@@ -60,6 +68,7 @@ def yellow_restock_cost(prices, stocks):
     提示：布林遮罩 + .sum()
     """
     # TODO: 你的程式碼
+    return (prices[prices < 500] * 50).sum()
     pass
 
 
@@ -77,4 +86,10 @@ def red_double11_prices(prices, stocks):
     提示：np.where 可以巢狀使用
     """
     # TODO: 你的程式碼
+    return np.where(
+        stocks >= 100, prices * 0.7,
+        np.where(
+            (stocks >=20) & (stocks <100), prices *0.9,
+            prices
+        ))
     pass

homework/m2_pandas_cleaning.py

@@ -20,6 +20,7 @@ def green_read_csv():
     提示：pd.read_csv()
     """
     # TODO: 你的程式碼
+    return pd.read_csv("../datasets/ecommerce/orders_raw.csv")
     pass
 
 
@@ -29,6 +30,8 @@ def green_shape(df):
     提示：df.shape
     """
     # TODO: 你的程式碼
+
+    return df.shape
     pass
 
 
@@ -38,6 +41,7 @@ def green_dtypes(df):
     提示：df.dtypes
     """
     # TODO: 你的程式碼
+    return df.dtypes
     pass
 
 
@@ -52,6 +56,9 @@ def yellow_clean_columns(df):
     提示：df.columns.str.strip().str.lower()
     """
     # TODO: 你的程式碼
+    df_copy = df.copy()
+    df_copy.columns = df_copy.columns.str.strip().str.lower()
+    return df_copy
     pass
 
 
@@ -63,6 +70,9 @@ def yellow_clean_amount(df):
     提示：.str.replace() + .astype(float)
     """
     # TODO: 你的程式碼
+    df_copy = df.copy()
+    df_copy["amount"] = df_copy["amount"].str.replace("$","",regex=False).str.replace(",","",regex=False).astype(float)
+    return df_copy
     pass
 
 
@@ -72,6 +82,7 @@ def yellow_drop_duplicates(df):
     提示：df.drop_duplicates()
     """
     # TODO: 你的程式碼
+    return df.drop_duplicates()
     pass
 
 
@@ -93,4 +104,11 @@ def red_clean_orders(path):
     提示：pd.to_datetime(errors='coerce')
     """
     # TODO: 你的程式碼
+    df = pd.read_csv(path)
+    df.columns = df.columns.str.strip().str.lower()
+    df["amount"] = df["amount"].str.replace("$","",regex=False).str.replace(",","",regex=False).astype(float)
+    df["order_date"] = pd.to_datetime(df["order_date"], errors='coerce')
+    df = df.dropna(subset=["amount","order_date"])
+    df = df.drop_duplicates()
+    return df
     pass

homework/m3_pandas_advanced.py

@@ -24,18 +24,27 @@ def green_load_and_merge():
     提示：pd.merge(how='left')
     """
     # TODO: 你的程式碼
+    df1 = pd.read_csv("../datasets/ecommerce/orders_clean.csv")
+    df2 = pd.read_csv("../datasets/ecommerce/customers.csv")
+    df3 = pd.read_csv("../datasets/ecommerce/products.csv")
+    df = df1.merge(df2, on = "customer_id", how = "left")
+    df = df.merge(df3, on = "product_id", how = "left")
+    return df
+
     pass
 
 
 def green_row_count(df):
     """回傳 DataFrame 的列數 (int)"""
     # TODO: 你的程式碼
+    return df.shape[0]
     pass
 
 
 def green_column_list(df):
     """回傳 DataFrame 的所有欄位名稱 (list)"""
     # TODO: 你的程式碼
+    return list(df.columns)
     pass
 
 
@@ -50,6 +59,7 @@ def yellow_top_category(df):
     提示：groupby('category')['amount'].sum()
     """
     # TODO: 你的程式碼
+    return df.groupby("category")["amount"].sum().sort_values().idxmax()
     pass
 
 
@@ -60,6 +70,9 @@ def yellow_gold_vip_stats(df):
     提示：df[df['vip_level'] == 'Gold']
     """
     # TODO: 你的程式碼
+    order_count = df[df["vip_level"] == "Gold"].shape[0]
+    total_amount = df[df["vip_level"] == "Gold"]["amount"].sum()
+    return (order_count, total_amount)
     pass
 
 
@@ -70,6 +83,7 @@ def yellow_region_avg_amount(df):
     提示：groupby('region')['amount'].mean()
     """
     # TODO: 你的程式碼
+    return df.groupby("region")["amount"].mean()
     pass
 
 
@@ -94,4 +108,15 @@ def red_rfm_top5(df):
     提示：groupby('customer_id').agg(...)
     """
     # TODO: 你的程式碼
+    result_df =  df.groupby("customer_id").agg({
+    "customer_name" : "first",
+    "order_date":"max",
+    "order_id" : "count",
+    "amount" : "sum"
+}).rename(columns={
+    "order_date" : "R",
+    "order_id" : "F",
+    "amount" : "M"
+}).sort_values("M", ascending= False).head().reset_index()
+    return result_df
     pass

homework/m4_timeseries.py

@@ -27,6 +27,8 @@ def green_avg_by_month():
     提示：df['order_date'].dt.month
     """
     # TODO: 你的程式碼
+    df = _load_data()
+    return df.groupby(df['order_date'].dt.month)["amount"].mean()
     pass
 
 
@@ -37,6 +39,8 @@ def green_top3_dates():
     提示：value_counts().head(3)
     """
     # TODO: 你的程式碼
+    df = _load_data()
+    return df["order_date"].value_counts().head(3)
     pass
 
 
@@ -46,6 +50,10 @@ def green_date_range():
     格式為 pandas Timestamp
     """
     # TODO: 你的程式碼
+    df = _load_data()
+    max_date = df["order_date"].max()
+    min_date = df["order_date"].min()
+    return (min_date, max_date)
     pass
 
 
@@ -60,6 +68,8 @@ def yellow_monthly_revenue():
     提示：set_index('order_date').resample('ME')['amount'].sum()
     """
     # TODO: 你的程式碼
+    df = _load_data()
+    return df.set_index(["order_date"]).resample("ME")["amount"].sum()
     pass
 
 
@@ -71,6 +81,7 @@ def yellow_rolling_avg(monthly_revenue):
     提示：.rolling(window=3).mean()
     """
     # TODO: 你的程式碼
+    return monthly_revenue.rolling(window=3).mean()
     pass
 
 
@@ -81,6 +92,7 @@ def yellow_category_median(df):
     提示：groupby + median + sort_values
     """
     # TODO: 你的程式碼
+    return df.groupby(df["category"])["amount"].median().sort_values(ascending=False)
     pass
 
 
@@ -101,4 +113,27 @@ def red_monthly_report():
     提示：resample + agg + pct_change
     """
     # TODO: 你的程式碼
+    df = _load_data()
+    monthly_report = (
+    df
+    .set_index('order_date')
+    .resample("ME")
+    .agg({
+    "order_id" : "count",
+    "amount" : "sum",
+    "customer_id" : "nunique"
+}).rename(columns={
+    "order_id" : "order_count",
+    "amount" : "revenue",
+    "customer_id" : "active_customers"
+})
+)
+
+    monthly_report.index = monthly_report.index.to_period("M")
+
+    monthly_report["avg_order_value"] = (monthly_report["revenue"]/monthly_report["order_count"])
+
+    monthly_report["revenue_growth"] = (monthly_report["revenue"].pct_change() * 100 )
+
+    return monthly_report
     pass

homework/m5_visualization.py

@@ -29,6 +29,10 @@ def green_bar_category():
     提示：sns.countplot 或 value_counts().plot.bar()
     """
     # TODO: 你的程式碼
+    df = _load_data()
+    plt.figure(figsize = (8, 4))
+    sns.countplot(data = df, x = "category")
+    return plt.gcf()
     pass
 
 
@@ -39,6 +43,10 @@ def green_hist_amount():
     提示：sns.histplot(bins=20) 或 plt.hist()
     """
     # TODO: 你的程式碼
+    df = _load_data()
+    plt.figure(figsize=(8,4))
+    sns.histplot(bins = 20, x = "amount", data = df)
+    return plt.gcf()
     pass
 
 
@@ -51,6 +59,13 @@ def green_set_labels():
     回傳 matplotlib Figure 物件
     """
     # TODO: 你的程式碼
+    df = _load_data()
+    plt.figure(figsize = (8, 4))
+    sns.countplot(data = df, x = "category")
+    plt.title("category_order_count")
+    plt.xlabel("category")
+    plt.ylabel("order_count")
+    return plt.gcf()
     pass
 
 
@@ -68,6 +83,21 @@ def yellow_line_region_trend():
     提示：分別 groupby 再 plot，或用 sns.lineplot(hue='region')
     """
     # TODO: 你的程式碼
+    df = _load_data()
+    df["month"] = df["order_date"].dt.to_period("M")
+    region_monthly_revenue = (
+        df[df["region"].isin(["North", "South"])]
+        .groupby(["month", "region"])["amount"]
+        .sum()
+        .unstack()
+    )
+    fig, ax = plt.subplots(figsize=(8, 4))
+
+    region_monthly_revenue.plot(
+        marker="o",
+        ax=ax
+    )
+    return fig
     pass
 
 
@@ -78,6 +108,10 @@ def yellow_box_vip():
     提示：sns.boxplot(x='vip_level', y='amount', data=df)
     """
     # TODO: 你的程式碼
+    df = _load_data()
+    fig, ax = plt.subplots(figsize=(8, 4))
+    sns.boxplot(x = "vip_level", y = "amount", data = df)
+    return fig
     pass
 
 
@@ -88,6 +122,10 @@ def yellow_scatter_price_amount():
     提示：plt.scatter() 或 sns.scatterplot()
     """
     # TODO: 你的程式碼
+    df = _load_data()
+    plt.figure(figsize=(8, 4))
+    plt.scatter(x = "unit_price", y = "amount", data = df)
+    return plt.gcf()
     pass
 
 
@@ -107,4 +145,21 @@ def red_category_dashboard(category="Electronics"):
     提示：fig, axes = plt.subplots(2, 2, figsize=(14, 10))
     """
     # TODO: 你的程式碼
+    df = _load_data()
+    df["month"] = df["order_date"].dt.to_period("M").astype("str")
+    df = df[df["category"].isin(["Electronics"])]
+    fig, axes = plt.subplots(2, 2, figsize = (14,10))
+
+
+    sns.lineplot(data=df, x = "month", y = "amount", marker="o", ax = axes[0, 0], estimator = "sum", errorbar=None )
+    axes[0, 0].tick_params(axis='x', rotation=45)
+
+    sns.barplot(data=df, x = "region", y = "amount", estimator="sum",
+        errorbar=None, ax = axes[0,1] )
+
+    df.groupby("product_name")["amount"].sum().sort_values(ascending=False).head(5).plot.barh(ax=axes[1,0])
+
+    sns.histplot(data = df, x = "amount", bins=20, ax=axes[1,1])
+
+    return fig
     pass