数据分析案例-汽车客户信息数据可视化分析（数据集+代码）.rar

# %% [markdown] # ID：客户ID # Gender：客户性别 # Ever_Married：客户婚姻状况 # Age：客户年龄 # Graduated：客户是毕业生吗？ # Profession：客户的职业 # Work_Experience：多年工作经验 # Spending_Score：客户的消费评分 # Family_Size：客户家庭成员人数（含客户） # Var_1：客户的匿名类别 # Segmentation：（目标）客户的客户群 # %% import numpy as np import pandas as pd import matplotlib.pyplot as plt from sklearn.cluster import KMeans import seaborn as sns from sklearn.cluster import DBSCAN df_train = pd.read_csv('Train.csv') df_train.head() # %% df_train.shape # %% df_train.info() # %% df_train.describe() # %% df_train.describe(include='O') # %% df_train.isnull().sum() # 统计缺失值 # %% df_train.dropna(inplace=True) # 删除缺失值 df_train.shape # %% any(df_train.duplicated()) #　检测数据集是否存在重复值 # %% # 不同组别的性别分布 sns.countplot(x='Segmentation', hue='Gender', data=df_train) plt.title("Segmentation based on Gender") plt.show() # %% # 不同组别的年龄分布 sns.boxplot(x='Segmentation', y='Age', data=df_train) plt.title("Age Distribution based on Segmentation") plt.show() # %% # 不同组别的工作经验vs消费得分 sns.boxplot(x='Work_Experience', y='Spending_Score', hue='Segmentation', data=df_train) plt.title("Work Experience vs Spending Score based on Segmentation") plt.show() # %% sns.violinplot(x='Segmentation', y='ID', data=df_train) plt.title("Profession Distribution based on Segmentation") plt.show() # %% # 不同组别的家庭规模 plt.figure(figsize=(8,8)) df_train['Family_Size'].value_counts().plot.pie(autopct='%1.1f%%') plt.title("Family Size Distribution based on Segmentation") plt.show() # %% # 客户年龄分布 plt.figure(figsize=(10,5)) sns.histplot(df_train['Age'], kde=True) plt.title("Distribution of 'Age'") plt.xlabel("Age") plt.ylabel("Frequency") plt.show() # %% # 工作经验分布 plt.figure(figsize=(10,5)) sns.histplot(df_train['Work_Experience'], kde=True) plt.title("Distribution of 'Work_Experience'") plt.xlabel("Work_Experience") plt.ylabel("Frequency") plt.show() # %% # 家庭规模分布 plt.figure(figsize=(10,5)) sns.histplot(df_train['Family_Size'], kde=True) plt.title("Distribution of 'Family_Size'") plt.xlabel("Family_Size") plt.ylabel("Frequency") plt.show() # %% # 删除目标变量 df_train_kmeans = df_train.drop(['Segmentation', 'ID'], axis=1) df_train_kmeans # %% # 将分类列转换为标签编码列 from sklearn.preprocessing import LabelEncoder encoder = LabelEncoder() df_train_kmeans['Gender'] = encoder.fit_transform(df_train_kmeans['Gender']) df_train_kmeans['Ever_Married'] = encoder.fit_transform(df_train_kmeans['Ever_Married']) df_train_kmeans['Graduated'] = encoder.fit_transform(df_train_kmeans['Graduated']) df_train_kmeans['Profession'] = encoder.fit_transform(df_train_kmeans['Profession']) df_train_kmeans['Spending_Score'] = encoder.fit_transform(df_train_kmeans['Spending_Score']) df_train_kmeans['Var_1'] = encoder.fit_transform(df_train_kmeans['Var_1']) df_train_kmeans.head() # %% # 相关系数矩阵 corr = df_train_kmeans.corr() # 绘制热力图 plt.figure(figsize=(10,10)) sns.heatmap(corr, annot=True) plt.show() # %% # 数据标准化 from sklearn.preprocessing import StandardScaler scaler = StandardScaler() df_train_kmeans_scaler = scaler.fit_transform(df_train_kmeans) df_train_kmeans_scaler # %% [markdown] # KMeans # %% # 肘部法寻找最佳簇数 Sum_of_squared_distances = [] K = range(2,15) for k in K: km = KMeans(n_clusters=k) km = km.fit(df_train_kmeans_scaler) Sum_of_squared_distances.append(km.inertia_) print("For k =", k, ", the inertia is", km.inertia_) # %% # 肘部法可视化 import matplotlib.pyplot as plt plt.plot(K, Sum_of_squared_distances, 'bx-') plt.xlabel('k') plt.ylabel('Sum_of_squared_distances') plt.title('Elbow Method For Optimal k') plt.show() # %% # 根据肘部法确定最佳聚类数 best_k = 4 # 训练KMeans模型 kmeans = KMeans(n_clusters=best_k) kmeans.fit(df_train_kmeans_scaler) # %% # 预测聚类标签 labels = kmeans.predict(df_train_kmeans_scaler) # 统计每个集群中的客户数量 from collections import Counter counts = Counter(labels) # 在“Segmentation”列中获得目标类别“A”、“B”、“C”和“D” target_categories = df_train['Segmentation'] # 找出每个簇中出现频率最高的目标类别 cluster_categories = {} for label, count in counts.items(): cluster_data = target_categories[labels == label] most_frequent_category = cluster_data.value_counts().idxmax() cluster_categories[label] = most_frequent_category # 将预测的标签映射到目标类别 mapped_labels = [cluster_categories[label] for label in labels] # 将列'n_Clusters'添加到具有映射标签的数据框 df_train['n_Clusters'] = mapped_labels df_train # %% # 计算正确预测的次数 correct_predictions = sum(df_train['Segmentation'] == df_train['n_Clusters']) # 计算预测的总数 total_predictions = df_train.shape[0] # 以百分比计算准确度 accuracy = (correct_predictions / total_predictions) * 100 print("-- Accuracy for KMeans: {:.2f}%".format(accuracy)) # %% [markdown] # DBSCAN # %% # 训练DBSCAN模型 dbscan = DBSCAN(eps=0.6, min_samples=5) dbscan.fit(df_train_kmeans_scaler) # 预测聚类标签 labels = dbscan.labels_ # 统计每个集群中的客户数量 from collections import Counter counts = Counter(labels) # 在` Segmentation `列中获得目标类别` A `， ` B `， ` C `和` D ` target_categories = df_train['Segmentation'] # 找出每个簇中出现频率最高的目标类别 cluster_categories = {} for label, count in counts.items(): cluster_data = target_categories[labels == label] most_frequent_category = cluster_data.value_counts().idxmax() cluster_categories[label] = most_frequent_category # 将预测的标签映射到目标类别 mapped_labels = [cluster_categories[label] if label in cluster_categories else 'Noise' for label in labels] # 将列'n_Clusters'添加到具有映射标签的数据框 df_train['n_Clusters'] = mapped_labels df_train # %% # 计算正确预测的次数 correct_predictions = sum(df_train['Segmentation'] == df_train['n_Clusters']) # 计算预测的总数 total_predictions = df_train.shape[0] # 以百分比计算准确度 accuracy = (correct_predictions / total_predictions) * 100 print("-- Accuracy for DBSCAN: {:.2f}%".format(accuracy))