如何将调查 pandas 数据框转换为可与 Python 中的 BI 工具一起使用的不同格式?

How to transform survey pandas dataframe into a different format usable with BI tools in Python?

我需要将调查结果转换成可在 BI 工具(如 Tableau)中使用的内容。

调查采用以下数据框格式

df = pd.DataFrame({'Respondent': ['Sally', 'Tony', 'Fred'],
               'What project did you work on with - Chris?': ['Project A','Project B', np.nan], 
               'What score would you give - Chris': [9,7,np.nan], 
               'Any other feedback for - Chris': ['Random Comment','Okay performance',np.nan],
               'What project did you work on with - Matt?': [np.nan,'Project C', 'Project X'], 
               'What score would you give - Matt': [np.nan,9,8], 
               'Any other feedback for - Matt': [np.nan, 'Great to work with Matt', 'Work was just okay'],
               'What project did you work on with - Luke?': ['Project B','Project D', 'Project Y'], 
               'What score would you give - Luke': [10,8,7], 
               'Any other feedback for - Luke': ['Work was excellent', 'Was a bit technical', 'Another Random Comment'],
              })

我需要将其转换为如下格式:

df = pd.DataFrame({'Name': ['Chris','Chris','Matt','Matt','Luke','Luke','Luke'],
               'Assessor': ['Sally','Tony','Tony','Fred','Sally','Tony','Fred'], 
               'Project Name': ['Project A', 'Project B', 'Project C', 'Project X', 'Project B', 'Project D', 'Project Y'], 
               'NPS Score': [9,7,9,8,10,8,7],
               'Feedback': ['Random Comment','Okay performance','Great to work with Matt','Work was just okay','Work was excellent','Was a bit technical','Another Random Comment']
              })

如您所见,它需要能够从列中提取名称。实际数据实际上要大得多,所以我需要代码可以处理任何大小,而不仅仅是这个例子。

new_data = pd.DataFrame(columns = ["Assessor", "Project Name","NPS Score","Feedback", "Name"])
i = 1
while i < (len(df.columns)):
    data = df.iloc[:,[0,i,i+1,i+2]]
    data["Name"] = str(data.columns[-1].split(" ")[-1])
    data.columns = ["Assessor", "Project Name","NPS Score","Feedback","Name"]
    new_data = new_data.append(data)
    i = i + 3
    
new_data = new_data.reset_index(drop = True)
new_data