某SLG游戏公测期用户与付费分析

#查看注册分布

reg_user &＃61; reg_user.groupby([‘register_time’]).user_id.count()

#画图

plt.style.use(‘dark_background’)

fig &＃61; plt.figure(figsize&＃61;(14, 10))

plt.plot(reg_user)

plt.title(‘用户注册分布图’)

plt.show()



#获取活跃用户(游戏时长大于30分钟小于24小时)

activate_user &＃61; data[(data[‘avg_online_minutes’]>&＃61;30) & (data[‘avg_online_minutes’]<&＃61;24*60)

#选取特征

activate_user_means &＃61; activate_user[[‘bd_stronghold_level’, ‘pvp_battle_count’, ‘pve_battle_count’, ‘avg_online_minutes’,

‘pay_count’, ‘pay_price’]]



#对玩家进行分类&＃xff0c;将特征近似的分成一类&＃xff0c;总共分成三类

estimator &＃61; KMeans(n_clusters&＃61;3)

res &＃61; estimator.fit_predict(activate_user_means)

#每一行数据对应的类

label_pred &＃61; estimator.labels_

#每一类的平均值

centroids&＃61;estimator.cluster_centers_



#将均值整合成DataFrame

means_data &＃61; {‘第一类’: centroids[0],

‘第二类’: centroids[1],

‘第三类’: centroids[2],}

activate_user_means &＃61; pd.DataFrame(means_data, index&＃61;[‘要塞等级’, ‘pvp次数’, ‘pve_次数’, ‘日均在线时间’,

‘付费次数’, ‘付费金额’])



#付费率&＃61;付费人数/活跃用户

pay_user &＃61; game[game[‘pay_price’]>0]

pay_rate &＃61; pay_user[‘user_id’].count() / act_user[‘user_id’].count()

print(‘付费率为%.2f’ %(pay_rate))



#ARPU &＃61; 付费金额/活跃人数

arpu &＃61; pay_user[‘pay_price’].sum() / act_user[‘user_id’].count()

print(‘arpu为%.2f元’ %(arpu))



#ARPPU &＃61; 付费金额/付费用户人数

arppu &＃61; pay_user[‘pay_price’].sum() / pay_user[‘user_id’].count()

print(‘arppu为%.2f元’ %(arppu))



#统计各个等级的玩家到达人数&＃xff0c;付费人数&＃xff0c; 付费次数&＃xff0c; 付费总额&＃xff0c; 等级付费转化率&＃xff0c;

等级人均付费次数&＃xff0c; 等级人均付费总额

user &＃61; data[[‘user_id’, ‘bd_stronghold_level’, ‘pay_count’, ‘pay_price’]]


#分成高级和低级玩家

eli_uesr &＃61; game[(game[‘pay_price’]>&＃61;500) & (game[‘bd_stronghold_level’]>&＃61;10)]

nor_user &＃61; game[(game[‘pay_price’]《500》) & (game[‘bd_stronghold_level’]>&＃61;10)]



#统计两类玩家各自的资源使用情况

wood_avg &＃61; [eli_uesr[‘wood_reduce_value’].mean(), nor_user[‘wood_reduce_value’].mean()]

stone_avg &＃61; [eli_uesr[‘stone_reduce_value’].mean(), nor_user[‘stone_reduce_value’].mean()]

ivory_avg &＃61; [eli_uesr[‘ivory_reduce_value’].mean(), nor_user[‘ivory_reduce_value’].mean()]

meat_avg &＃61; [eli_uesr[‘meat_reduce_value’].mean(), nor_user[‘meat_reduce_value’].mean()]

magic_avg &＃61; [eli_uesr[‘magic_reduce_value’].mean(), nor_user[‘magic_reduce_value’].mean()]



data &＃61; {‘高端玩家’:[wood_avg[0], stone_avg[0], ivory_avg[0], meat_avg[0], magic_avg[0]],

‘低端玩家’:[wood_avg[1], stone_avg[1], ivory_avg[1], meat_avg[1], magic_avg[1]]}

resource &＃61; pd.DataFrame(data, index&＃61;[‘木头’, ‘石头’, ‘象牙’, ‘肉’, ‘魔法’])


#各类玩家对加速券的使用情况

genral_avg &＃61; [eli_uesr[‘general_acceleration_reduce_value’].mean(), nor_user[‘general_acceleration_reduce_value’].mean()]

building_avg &＃61; [eli_uesr[‘building_acceleration_reduce_value’].mean(), nor_user[‘building_acceleration_reduce_value’].mean()]

reaserch_avg &＃61; [eli_uesr[‘reaserch_acceleration_reduce_value’].mean(), nor_user[‘reaserch_acceleration_reduce_value’].mean()]

training_avg &＃61; [eli_uesr[‘training_acceleration_reduce_value’].mean(), nor_user[‘training_acceleration_reduce_value’].mean()]

treament_avg &＃61; [eli_uesr[‘treatment_acceleration_reduce_value’].mean(), nor_user[‘treatment_acceleration_reduce_value’].mean()]



data &＃61; {‘高端玩家’:[genral_avg[0], building_avg[0], reaserch_avg[0], training_avg[0], treament_avg[0]],

‘低端玩家’:[genral_avg[1], building_avg[1], reaserch_avg[1], training_avg[1], treament_avg[1]]}



acceleration &＃61; pd.DataFrame(data, index &＃61; [‘通用’, ‘建筑’, ‘科研’, ‘训练’, ‘治疗’])

#绘图

plt.style.use(‘dark_background’)

acceleration.plot(kind &＃61; ‘bar’, stacked&＃61;True, colormap&＃61; ‘Reds’, figsize&＃61;(14, 10))

plt.title(‘玩家加速券使用量’)

plt.show()