pandas实战教程1_热门电影

1 简介

刚学完pandas基础后跟着老师写的一个热门电影数据分析，电影分析数据集地址

2 过程

1. 导入数据

unames = ["user_id", "gender", "age", "occupation", "zip"]
users = pd.read_table("ml-1m/users.dat", sep="::", header=None, names=unames)


rating_names = ["user_id", "movie_id", "rating", "timestamp"]
ratings = pd.read_table("ml-1m/ratings.dat", names=rating_names, sep="::", header=None)
ratings.head(5)

movies_names = ["movie_id", "title", "genrez"]
movies = pd.read_table("ml-1m/movies.dat", names=movies_names, sep="::", header=None)
movies.head(5)   

• 写出一个数据的columns 为names赋值
• sep 表示分隔符是哪个
• header=None 没有头
  
  

2.数据连接 merge

data = pd.merge(pd.merge(users, ratings), movies)

• 两次merge 这里直接通过下标merge了
  

3.数据透视表 筛选数据

ratings_by_gender = data.pivot_table(values="rating", index="title", columns="gender", aggfunc="mean")

• values 数值
• index 行索引
• columns 列索引
• aggfun 处理函数
  

4.寻找热门电影

1. 查出评论最多的几部电影 groupby

ratings_by_title = data.groupby("title").size()
hot_movies = ratings_by_title[ratings_by_title > 1000]

2. 查出平均得分

mean_ratings = data.pivot_table(index="title", values="rating",aggfunc="mean")
mean_ratings_series = pd.Series(mean_ratings["rating"]) #在这一步遇到麻烦，发现由数据透视表的得到的是DataFrame要选择的话需要转成Series
hot_movies_rating = mean_ratings_series[hot_movies.index]
top_10_movies = hot_movies_rating.sort_values(ascending=False).head(10)