对duration进行精度为0.2的预处理

df['duration'] = df['duration'].apply(lambda x: round(x/0.2)*0.2)

统计duration和subscribe的关系

duration_subscribe = df.groupby('duration')['subscribe'].mean()

生成可视化图

p = figure(title='Duration vs Subscribe', x_axis_label='Duration', y_axis_label='Subscribe Rate') p.line(duration_subscribe.index, duration_subscribe.values) output_file('duration_subscribe.html') show(p)