hadoop学习---基于Hive的聊天数据分析报表可视化案例
背景介绍:
聊天平台每天都会有大量的用户在线,会出现大量的聊天数据,通过对聊天数据的统计分析,可以更好的对用户构建精准的用户画像,为用户提供更好的服务以及实现高ROI的平台运营推广,给公司的发展决策提供精确的数据支撑。 我们将基于一个社交平台App的用户数据,完成相关指标的统计分析并结合BI工具对指标进行可视化展现。
目标:
基于Hadoop和Hive实现聊天数据统计分析,构建聊天数据分析报表
需求分析:
统计今日总消息量 统计今日每小时消息量、发送和接收用户数
统计今日各地区发送消息数据量
统计今日发送消息和接收消息的用户数
统计今日发送消息最多的Top10用户
统计今日接收消息最多的Top10用户
统计发送人的手机型号分布情况
统计发送人的设备操作系统分布情况
数据源:数据源
数据大小:30万条数据
列分隔符:Hive默认分隔符’\001’
数据字典及样例数据
构建数据库数据表hive:
create database db_msg; use db_msg; --建表 create table db_msg.tb_msg_source( msg_time string comment "消息发送时间", sender_name string comment "发送人昵称", sender_account string comment "发送人账号", sender_sex string comment "发送人性别", sender_ip string comment "发送人ip地址", sender_os string comment "发送人操作系统", sender_phonetype string comment "发送人手机型号", sender_network string comment "发送人网络类型", sender_gps string comment "发送人的GPS定位", receiver_name string comment "接收人昵称", receiver_ip string comment "接收人IP", receiver_account string comment "接收人账号", receiver_os string comment "接收人操作系统", receiver_phonetype string comment "接收人手机型号", receiver_network string comment "接收人网络类型", receiver_gps string comment "接收人的GPS定位", receiver_sex string comment "接收人性别", msg_type string comment "消息类型", distance string comment "双方距离", message string comment "消息内容" );
将数据源上传到Linux文件系统中,再上传到hadoop中:
[hadoop@node1 ~]$ hadoop fs -mkdir -p /chatdemo/data [hadoop@node1 ~]$ hadoop fs -put chat_data-30W.csv /chatdemo/data
将数据源从hadoop中下载到hive数据库中:
load data inpath '/chatdemo/data/chat_data-30W.csv' into table db_msg.tb_msg_source;
数据清洗:
问题1:当前数据中,有一些数据的字段为空,不是合法数据
问题2:需求中,需要统计每天、每个小时的消息量,但是数据中没有天和小时字段,只有整体时间字段,不好处理
问题3:需求中,需要对经度和维度构建地区的可视化地图,但是数据中GPS经纬度为一个字段,不好处理
数据清洗:
select *, date(msg_time) as msg_day, hour(msg_time) as msg_hour, split(sender_gps,',')[0] as sender_lng, split(sender_gps,',')[1] as sender_lat from tb_msg_source where length(sender_gps)>0;
将清洗后的数据在放入新表中,添加新的字段:
create table db_msg.tb_msg_etl( msg_time string comment "消息发送时间", sender_name string comment "发送人昵称", sender_account string comment "发送人账号", sender_sex string comment "发送人性别", sender_ip string comment "发送人ip地址", sender_os string comment "发送人操作系统", sender_phonetype string comment "发送人手机型号", sender_network string comment "发送人网络类型", sender_gps string comment "发送人的GPS定位", receiver_name string comment "接收人昵称", receiver_ip string comment "接收人IP", receiver_account string comment "接收人账号", receiver_os string comment "接收人操作系统", receiver_phonetype string comment "接收人手机型号", receiver_network string comment "接收人网络类型", receiver_gps string comment "接收人的GPS定位", receiver_sex string comment "接收人性别", msg_type string comment "消息类型", distance string comment "双方距离", message string comment "消息内容", msg_day string comment "消息日", msg_hour string comment "消息小时", sender_lng double comment "经度", sender_lat double comment "纬度" );
将数据插入到新表:
insert overwrite table db_msg.tb_msg_etl select *, date(msg_time) as msg_day, hour(msg_time) as msg_hour, split(sender_gps,',')[0] as sender_lng, split(sender_gps,',')[1] as sender_lat from tb_msg_source where length(sender_gps)>0;
将需求指标查询出来构建新表:
-- 统计今日总消息量 create table db_msg.tb_rs_total_msg_cnt comment '每日消息总量' as select msg_day, count(*) as total_msg_cnt from db_msg.tb_msg_etl group by msg_day ; -- 统计今日每小时消息量、发送和接收用户数 create table db_msg.tb_rs_hour_msg_cnt comment '每小时消息量趋势' as select msg_hour, count(*) as total_msg_cnt, count(distinct sender_account) as sender_user_cnt, count(distinct receiver_account) as receiver_user_cnt from db_msg.tb_msg_etl group by msg_hour; -- 统计今日各地区发送消息数据量 create table db_msg.tb_rs_loc_cnt comment '每日各地区发送消息总量' as select msg_day,sender_lng,sender_lat,count(*) as total_msg_cnt from db_msg.tb_msg_etl group by msg_day,sender_lng,sender_lat -- 统计今日发送消息和接收消息的用户数 create table db_msg.tb_rs_user_cnt comment '每日发送和接收消息的人数' as select msg_day, count(distinct sender_account) as sender_user_cnt, count(distinct receiver_account) as receiver_user_cnt from db_msg.tb_msg_etl group by msg_day; -- 统计今日发送消息最多的Top10用户 create table db_msg.tb_rs_s_user_top10 comment '发送消息最多的10个用户' as select sender_name, count(*) as sender_msg_cnt from db_msg.tb_msg_etl group by sender_name order by sender_msg_cnt desc limit 10; -- 统计今日接收消息最多的Top10用户 create table db_msg.tb_rs_r_user_top10 comment '接受消息最多的10个用户' as select receiver_name, count(*) as receiver_msg_cnt from db_msg.tb_msg_etl group by tb_msg_etl.receiver_name order by receiver_msg_cnt desc limit 10; -- 统计发送人的手机型号分布情况 create table db_msg.tb_rs_sender_phone comment '发送人的手机型号分布' as select sender_phonetype, count(*) as cnt from db_msg.tb_msg_etl group by sender_phonetype; -- 统计发送人的设备操作系统分布情况 create table db_msg.tb_rs_sender_os comment '发送人的os分布' as select sender_os, count(*) as cnt from db_msg.tb_msg_etl group by sender_os;
基于FineBI完成指标的可视化展示:
文章版权声明:除非注明,否则均为主机测评原创文章,转载或复制请以超链接形式并注明出处。