Compared to The first version of Douyu_danmu, in this repository, the analysis of Douyu_TV's danmu is based on SPARK instead of MYSQL(Pymysql).

Python 3.6
Module jieba, wordcloud
Spyder
SPARK (Pyspark)
Windows10 (64bit)

In Anaconda Prompt/CMD, print"python Spark_danmu_scrapy.py", and then input the room-id to activate the scrapy process.
OR use the .exe app in the link below.

After the live-broadcast show, stop the scrapy process. In Anaconda Prompt/CMD, print"python Spark_danmu_analyze.py", and then input the room-id to activate the analyze process.

The results include: Hot Words/The histogram of level/The Top5 badges and so on. Just as shown in
and

(two examples)

The psd files are the templets that I use to make the daily reports.
Like nvliu66 and yjjimpaopao.Like 156277 and 687423 (ง•̀_•́)ง

BAIDU CLOUD: 链接(Link): http://t.cn/R8MZkGV 密码(PWD): h5ed

Monthly or Yearly Report——By applying KMEANS to help host improve the LIVE.