liubiren
/
Python
1
1
Fork 0
Code Issues Pull Requests Packages Projects Releases Activity

251111更新

This commit is contained in:
marslbr 2025-11-11 17:33:22 +08:00
parent 2fedbac2a5
commit 16a0b4e7c5
2 changed files with 121 additions and 436 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

557
KANO/main.py
View File

@ -5,9 +5,12 @@
本脚本用于KANO需求分析，能够处理问卷数据并输出需求分类结果 本脚本用于KANO需求分析，能够处理问卷数据并输出需求分类结果
""" """
import re
import numpy
import pandas import pandas
import prettytable from utils.pandas_extension import save_as_workbook
print("1 打开并读取Excel文件...", end="") print("1 打开并读取Excel文件...", end="")
@ -37,9 +40,9 @@ try:
raise Exception("样本数为0") raise Exception("样本数为0")
# 统计列数 # 统计列数
columns_counts = dataset.shape[1] columns = dataset.shape[1]
# 若列数非奇数则抛出异常 # 若列数非奇数则抛出异常
if columns_counts % 2 != 0: if columns % 2 != 0:
raise Exception("列数为奇数") raise Exception("列数为奇数")
print(f"已完成，样本数为{samples_size}") print(f"已完成，样本数为{samples_size}")
@ -48,451 +51,133 @@ except Exception as exception:
print(f"发生异常：{str(exception)}") print(f"发生异常：{str(exception)}")
exit() exit()
# 读取问卷题目和答案（从第7列开始为题目或答案） print("2 生成KANO评价结果...", end="")
DataSet = DataSet.iloc[:, 6:]
# 统计数据集中样本数量和题目数量 # KANO评价结果
Sample_Size, Question_Amount = DataSet.shape kano = pandas.DataFrame()
# 判断题目数量是否为偶数，若为偶数则计算问卷中涉及需求数量，若为奇数则终止脚本 for column in range(columns):
if Question_Amount % 2 == 0: if column % 2 != 0:
continue
# 计算问卷中涉及需求数量 # 功能名称
Requirement_Amount = int(Question_Amount / 2) feature = re.search(
pattern="【(?P<feature>.*?)】", string=dataset.columns[column]
).group("feature")
else: # 生成某功能的选项分布
distribution = pandas.crosstab(
index=dataset.iloc[:, column], # 行索引为如果该产品有{{功能名称}}的选项
columns=dataset.iloc[:, column + 1], # 列索引为如果该产品没有{{功能名称}}的选项
rownames=[""],
colnames=[""],
normalize=True, # 转为占比
).reindex(index=alternatives, columns=alternatives, fill_value=0)
print("题目数量为奇数，请检查") kano = pandas.concat(
print("") objs=[
kano,
exit() pandas.DataFrame(
[
print( {
"数据集中包含 %d 份样本， %d 个问题（涉及 %d 个需求）" "功能名称": feature,
% (Sample_Size, Question_Amount, Requirement_Amount) "魅力属性": distribution.loc[
) "非常满意", ["理应如此", "无所谓", "勉强接受"]
print("") ].sum(),
"期望属性": distribution.loc["非常满意", "不满意"].sum(),
print("*" * 100) "必备属性": distribution.loc[
print("") ["理应如此", "无所谓", "勉强接受"], "不满意"
].sum(),
print("2 数据预处理") "无差异属性": distribution.loc[
print("") ["理应如此", "无所谓", "勉强接受"],
["理应如此", "无所谓", "勉强接受"],
print("2.1 检查并删除不规范样本") ]
print("") .sum()
.sum(),
# 定义问卷中备选答案列表 "反向属性": distribution.loc[
Alternatives = ["我很喜欢", "理所应当", "无所谓", "勉强接受", "我很不喜欢"] ["理应如此", "无所谓", "勉强接受", "不满意"], "非常满意"
].sum()
# 检查答案是否在指定范围，若否则将该答案定义为空值 + distribution.loc[
DataSet = DataSet.applymap(lambda x: x if x in Alternatives else None) ("不满意", ["理应如此", "无所谓", "勉强接受"])
].sum(),
# 删除包含缺失值的样本 "可疑结果": distribution.loc["非常满意", "非常满意"]
DataSet.dropna(axis="index", how="any", inplace=True) + distribution.loc["不满意", "不满意"],
}
# 删除答案全部相同的样本 ]
DataSet = DataSet[DataSet.apply(pandas.Series.nunique, axis="columns") != 1]
# 统计有效样本数量
Sample_Size = DataSet.shape[0]
print("处理后，有效样本数量为 %d 份" % (Sample_Size))
print("")
print("*" * 100)
print("")
print("3 数据处理")
print("")
Requirement_Labels = DataSet.columns.tolist()
# 通过问题截取需求名称（截取'有'右侧、'，'左侧字符串）
Requirement_Labels = [
x[x.find("有") + 1 : x.find("，")]
for x in Requirement_Labels
if isinstance(x, str) and "具有" in x
]
print("3.1 绘制KANO评价结果分类对照表")
print("")
for Question_Number in range(Requirement_Amount):
# 创建KANO评价结果分类对照表
KANO = pandas.DataFrame(data=[], index=Alternatives, columns=Alternatives)
for Column_Label in Alternatives:
for Index_Label in Alternatives:
# 统计并赋值
KANO.loc[Index_Label, Column_Label] = DataSet.loc[
(DataSet.iloc[:, Question_Number].isin([Index_Label]))
& (DataSet.iloc[:, Question_Number + 1].isin([Column_Label])),
:,
].shape[0]
# 修改行名
KANO.index = [
"Provide_Like",
"Provide_Should",
"Provide_Indifferent",
"Provide_Grudging",
"Provide_Hate",
]
# 重置索引
KANO.reset_index(inplace=True)
# 修改列名
KANO.columns = [
"",
"Not_Provide_Like",
"Not_Provide_Should",
"Not_Provide_Indifferent",
"Not_Provide_Grudging",
"Not_Provide_Hate",
]
# 打印表格
PrintTable = prettytable.PrettyTable()
PrintTable.field_names = KANO.columns.tolist()
for Index in KANO.index.tolist():
PrintTable.add_row(KANO.loc[Index])
PrintTable.align = "r"
PrintTable.align[""] = "l"
PrintTable.float_format = "."
print(
"附表 需求%d：%s的KANO评价结果分类对照表："
% (Question_Number + 1, Requirement_Labels[Question_Number])
)
print(PrintTable)
print("")
print("字段说明：")
print(
"1）Not_Provide_Like为不提供该需求、用户表示“我很喜欢”，Not_Provide_Should为不提供该需求、用户表示“理所应当”，Not_Provide_Indifferent为不提供该需求、用户表示“无所谓”，Not_Provide_Grudging为不提供该需求、用户表示“勉强接受”，Not_Provide_Hate为不提供该需求、用户表示“我很不喜欢”。"
)
print(
"1）Provide_Like为提供该需求、用户表示“我很喜欢”，Provide_Should为提供该需求、用户表示“理所应当”，Provide_Indifferent为提供该需求、用户表示“无所谓”，Provide_Grudging为提供该需求、用户表示“勉强接受”，Provide_Hate为不提供该需求、用户表示“我很不喜欢”。"
)
print("")
print("3.2 计算KANO评价维度")
print("")
# 创建KANO各维度分数表
KANO_Report = pandas.DataFrame(
data=[],
columns=[
"Requirement_Label",
"A_Score",
"O_Score",
"M_Score",
"I_Score",
"R_Score",
"Q_Score",
],
dtype="float",
)
KANO_Report["Requirement_Label"] = Requirement_Labels
for Question_Number in range(Requirement_Amount):
# 计算兴奋型需求分数
A_Score = round(
DataSet.loc[
(DataSet.iloc[:, Question_Number].isin(["我很喜欢"]))
& (
DataSet.iloc[:, Question_Number + 1].isin(
["理所应当", "无所谓", "勉强接受"]
)
), ),
:, ],
].shape[0] ignore_index=True,
/ Sample_Size
* 100,
2,
) )
KANO_Report.loc[Question_Number, "A_Score"] = A_Score # 计算better-worse分数，并贴标签
kano = (
# 计算期望型需求分数 kano.assign(
O_Score = round( 满意系数=lambda dataframe: (dataframe["魅力属性"] + dataframe["期望属性"])
DataSet.loc[ / (
(DataSet.iloc[:, Question_Number].isin(["我很喜欢"])) dataframe["魅力属性"]
& (DataSet.iloc[:, Question_Number + 1].isin(["我很不喜欢"])), + dataframe["期望属性"]
:, + dataframe["必备属性"]
].shape[0] + dataframe["无差异属性"]
/ Sample_Size ),
* 100, 不满意系数=lambda dataframe: (dataframe["必备属性"] + dataframe["期望属性"])
2, / (
dataframe["魅力属性"]
+ dataframe["期望属性"]
+ dataframe["必备属性"]
+ dataframe["无差异属性"]
),
) )
.pipe(
KANO_Report.loc[Question_Number, "O_Score"] = O_Score lambda dataframe: dataframe.assign(
需求标签=numpy.select(
# 计算必备型需求分数 condlist=[
M_Score = round( (dataframe["满意系数"] >= dataframe["满意系数"].mean())
DataSet.loc[ & (dataframe["不满意系数"] >= dataframe["不满意系数"].mean()),
(DataSet.iloc[:, Question_Number].isin(["理所应当", "无所谓", "勉强接受"])) (dataframe["满意系数"] >= dataframe["满意系数"].mean())
& (DataSet.iloc[:, Question_Number + 1].isin(["我很不喜欢"])), & (dataframe["不满意系数"] < dataframe["不满意系数"].mean()),
:, (dataframe["满意系数"] < dataframe["满意系数"].mean())
].shape[0] & (dataframe["不满意系数"] < dataframe["不满意系数"].mean()),
/ Sample_Size (dataframe["满意系数"] < dataframe["满意系数"].mean())
* 100, & (dataframe["不满意系数"] >= dataframe["不满意系数"].mean()),
2, ],
) choicelist=[
"期望型需求(P2)",
KANO_Report.loc[Question_Number, "M_Score"] = M_Score "魅力型需求(P3)",
"无差异需求(P4)",
# 计算无差型需求分数 "必备型需求(P1)",
I_Score = round( ],
DataSet.loc[ default="未分类",
(DataSet.iloc[:, Question_Number].isin(["理所应当", "无所谓", "勉强接受"]))
& (
DataSet.iloc[:, Question_Number + 1].isin(
["理所应当", "无所谓", "勉强接受"]
)
), ),
:, )
].shape[0]
/ Sample_Size
* 100,
2,
) )
.assign(
KANO_Report.loc[Question_Number, "I_Score"] = I_Score 魅力属性=lambda dataframe: dataframe["魅力属性"].apply(
lambda cell: f"{cell:.2f}"
# 计算反向型需求分数 ),
R_Score = round( 期望属性=lambda dataframe: dataframe["期望属性"].apply(
DataSet.loc[ lambda cell: f"{cell:.2f}"
( ),
DataSet.iloc[:, Question_Number].isin( 必备属性=lambda dataframe: dataframe["必备属性"].apply(
["理所应当", "无所谓", "勉强接受", "我很不喜欢"] lambda cell: f"{cell:.2f}"
) ),
) 无差异属性=lambda dataframe: dataframe["无差异属性"].apply(
& ( lambda cell: f"{cell:.2f}"
DataSet.iloc[:, Question_Number + 1].isin( ),
["我很喜欢", "理所应当", "无所谓", "勉强接受"] 反向属性=lambda dataframe: dataframe["反向属性"].apply(
) lambda cell: f"{cell:.2f}"
), ),
:, 可疑结果=lambda dataframe: dataframe["可疑结果"].apply(
].shape[0] lambda cell: f"{cell:.2f}"
/ Sample_Size ),
* 100 满意系数=lambda dataframe: dataframe["满意系数"].apply(
- I_Score, lambda cell: f"{cell:.2f}"
2, ),
) 不满意系数=lambda dataframe: dataframe["不满意系数"].apply(
lambda cell: f"{cell:.2f}"
KANO_Report.loc[Question_Number, "R_Score"] = R_Score ),
# 计算可疑型需求分数
Q_Score = round(
DataSet.loc[
(DataSet.iloc[:, Question_Number].isin(["我很喜欢"]))
& (DataSet.iloc[:, Question_Number + 1].isin(["我很喜欢"])),
:,
].shape[0]
/ Sample_Size
* 100
+ DataSet.loc[
(DataSet.iloc[:, Question_Number].isin(["我很不喜欢"]))
& (DataSet.iloc[:, Question_Number + 1].isin(["我很不喜欢"])),
:,
].shape[0]
/ Sample_Size
* 100,
2,
)
KANO_Report.loc[Question_Number, "Q_Score"] = Q_Score
# 打印表格
PrintTable = prettytable.PrettyTable()
PrintTable.field_names = KANO_Report.columns.tolist()
for Index in KANO_Report.index.tolist():
PrintTable.add_row(KANO_Report.loc[Index])
PrintTable.align["Requirement_Label"] = "l"
PrintTable.align["A_Score"] = "r"
PrintTable.align["O_Score"] = "r"
PrintTable.align["M_Score"] = "r"
PrintTable.align["I_Score"] = "r"
PrintTable.align["R_Score"] = "r"
PrintTable.align["Q_Score"] = "r"
PrintTable.align["Better_Score"] = "r"
PrintTable.align["Worse_Score"] = "r"
PrintTable.float_format["A_Score"] = ".2"
PrintTable.float_format["O_Score"] = ".2"
PrintTable.float_format["M_Score"] = ".2"
PrintTable.float_format["I_Score"] = ".2"
PrintTable.float_format["R_Score"] = ".2"
PrintTable.float_format["Q_Score"] = ".2"
PrintTable.float_format["Better_Score"] = ".2"
PrintTable.float_format["Worse_Score"] = ".2"
print("附表 各需求的KANO评价维度计算结果：")
print(
PrintTable.get_string(
fields=[
"Requirement_Label",
"A_Score",
"O_Score",
"M_Score",
"I_Score",
"R_Score",
"Q_Score",
]
) )
) )
print("字段说明：") save_as_workbook(worksheets=[("Sheet1", kano)], workbook_name="result.xlsx")
print( print("已完成")
"1）Requirement_Label为需求名称，A_Score为兴奋型需求分数，O_Score为期望型需求分数，M_Score为必备型需求分数，I_Score为无差型需求分数，R_Score为反向型需求分数，Q_Score为可疑型需求分数。"
)
print("")
print("3.3 定义需求类型和Better-Worse分数")
print("")
# 以KANO评价维度中最高分定义需求类型
Requirement_Types = list(
KANO_Report[
["A_Score", "O_Score", "M_Score", "I_Score", "R_Score", "Q_Score"]
].idxmax(axis="columns")
)
# 通过列名截取需求类型（第一位、'_'左侧字符串）
Requirement_Types = [
x[0 : x.find("_")] for x in Requirement_Types if isinstance(x, str)
]
KANO_Report["Requirement_Type"] = Requirement_Types
# 计算Better分数
KANO_Report["Better_Score"] = (
(KANO_Report["A_Score"] + KANO_Report["O_Score"])
/ (
KANO_Report["A_Score"]
+ KANO_Report["O_Score"]
+ KANO_Report["M_Score"]
+ KANO_Report["I_Score"]
)
* 100
)
# 计算Worse分数
KANO_Report["Worse_Score"] = (
-1
* (KANO_Report["O_Score"] + KANO_Report["M_Score"])
/ (
KANO_Report["A_Score"]
+ KANO_Report["O_Score"]
+ KANO_Report["M_Score"]
+ KANO_Report["I_Score"]
)
* 100
)
# 打印表格
PrintTable = prettytable.PrettyTable()
PrintTable.field_names = KANO_Report.columns.tolist()
for Index in KANO_Report.index.tolist():
PrintTable.add_row(KANO_Report.loc[Index])
PrintTable.align["Requirement_Label"] = "l"
PrintTable.align["Requirement_Type"] = "r"
PrintTable.align["Better_Score"] = "r"
PrintTable.align["Worse_Score"] = "r"
PrintTable.float_format["Better_Score"] = ".2"
PrintTable.float_format["Worse_Score"] = ".2"
print("附表 各需求的KANO评价维度计算结果：")
print(
PrintTable.get_string(
fields=["Requirement_Label", "Requirement_Type", "Better_Score", "Worse_Score"]
)
)
print("字段说明：")
print("1）Requirement_Label为需求名称，Requirement_Type为需求类型。")
print(
"2）A为兴奋型需求：表示产品具有该种需求，则用户满意度会提高；没有该种需求，则用户满意度不会降低。建议给予P3关注。"
)
print(
"3）O为期望型需求：表示产品具有该种需求，则用户满意度会提高；没有该种需求，则用户满意度会降低。建议给予P1关注。"
)
print(
"4）M为必备型需求：表示产品具有该种需求，则用户满意度不会提高；没有该种需求，则用户满意度会降低。建议给予P2关注。"
)
print(
"5）I为无差型需求：表示产品具有该种需求，则用户满意度不会提高；没有该种需求，则用户满意度不会降低。建议给予P4关注。"
)
print("6）R为反向型需求：建议给予关注。")
print("7）Q为可疑型需求：建议给予关注。")
print(
"8）Better_Score为Better分数。表示如果产品具有某种需求，则用户满意度会提高，数值越大提高越大。"
)
print(
"9）Worse_Score为Worse分数。表示如果产品没有某种需求，则用户满意度会下降，数值越小下降越大。"
)
print("")

BIN
KANO/result.xlsx Normal file
View File

Binary file not shown.