作为背景,我试图使用回归来衡量某一周内竞争对手广告的出现是否会影响该广告的衡量标准.我不确定如何合并周,或根据品牌在一周内的存在情况(但在不同的行上)分配布尔值(1或0).
import pandas as pd
df = pd.DataFrame({'week': ['2019-11-11', '2019-11-11', '2019-11-18', '2019-11-25', '2019-11-11', '2019-11-18', '2019-11-11'],
'brand':['X', 'X-2', 'X', 'X', 'Y', 'Y', 'Z'],
'score': [.34, .25, .54, .23, .22, .34, .44]})
预期结果:
df = pd.DataFrame({'week': ['2019-11-11', '2019-11-11', '2019-11-18', '2019-11-25', '2019-11-11', '2019-11-18', '2019-11-11'],
'brand':['X', 'X-2', 'X', 'X', 'Y', 'Y', 'Z'],
'score': [.34, .25, .54, .23, .22, .34, .44],
'presence_dummy_Y': [1, 1, 1, 0, 1, 1, 1],
'presence_dummy_Z': [1, 1, 0, 0, 1, 0, 1]})