#  read

tips = pd.read_csv('d:/tips.csv', encoding='UTF-8')
unicode_minus = False # 解决保存图像是负号'-'显示为方块的问题,或者转换负号为字符串
print("1.2题:n", tips)

df1 = pd.read_csv('d:/tips.csv', encoding='UTF-8', nrows=5)
print("1.3题:n", df1)

print("1.4题:n")
df = pd.read_csv('d:/tips.csv', encoding='UTF-8')
del df['Unnamed: 0']
print('Unnamed:0列删除后:n', df)

print("1.5题:n", tips.iloc[[0, 2, 3]])

print("1.6题:n")
df = pd.DataFrame(tips)
print('describe=n',df.describe())

print("1.7题:n", tips.groupby(['day', 'time']).get_group(('Thur', 'Dinner')))

print("1.8题:n", tips.groupby('sex').agg([np.mean, np.std]))

print("1.9题:n", tips.pivot_table(values='total_bill', index='time', columns='day', margins=True, margins_name='合计', fill_value=0))

print("1.10题:n")
total_bill = [16.99, 10.34, 23.68, 24.29, 25.29, 8.77, 26.88]
tip=[1.01, 1.66, 3.31, 3.61, 4.71, 2, 3.12]
linear = interpolate.interp1d(total_bill,tip,kind='linear')
plt.plot(linear([16.99, 10.34, 21.01, 23.68, 24.59, 25.29, 8.77, 26.88]), '-.')
print(linear(21.01))
plt.show()
tips = df.fillna(linear(21.01))# 将求出值插入后的新tips
df = pd.DataFrame(tips)
print(df.iloc[[0, 1, 2, 3, 4]])

print("1.11题:n")
s = np.random.rand(20)
cuts = pd.cut(s, 4)
# print(cuts)
print(cuts.codes)
print(pd.value_counts(cuts))

print("1.12题:n")
plt.subplot(2, 2, 1)
data = tips['total_bill']
plt.hist(data, facecolor="blue", edgecolor="black", alpha=0.7)
plt.xlabel("区间", fontproperties='SimHei')
plt.ylabel("频数/频率", fontproperties='SimHei')
plt.title("子图1.Total bill", fontproperties='SimHei')
# plt.show()

print("1.13题:n")
plt.subplot(2, 2, 2)
x = tips['total_bill']
y = tips['tip']
plt.rcParams['font.sans-serif'] = ['SimHei']#黑体字体
plt.rcParams['axes.unicode_minus'] = False # 解决保存图像是负号'-'显示为方块的问题,或者转换负号为字符串
plt.scatter(x, y, c='green')
plt.xlabel("total_bill")
plt.ylabel("tip")
plt.title("子图2.Total bill--tip")
# plt.show()

print("1.14题:n")
plt.subplot(2, 2, 3)
plt.rcParams['font.sans-serif'] = ['SimHei']#黑体字体
plt.rcParams['axes.unicode_minus'] = False # 解决保存图像是负号'-'显示为方块的问题,或者转换负号为字符串
groups = df.groupby(['time']).get_group(('Dinner'))
x = groups['size']
groups = df.groupby(['time']).get_group(('Dinner'))
y = groups['total_bill']
plt.scatter(x, y, c='red')
plt.xlabel("size")
plt.ylabel("total_bill")
plt.title("子图3.Size--total_bill")
# plt.show()

print("1.15题:n")
plt.subplot(2, 2, 4)
plt.rcParams['font.sans-serif'] = ['SimHei'] # 黑体字体
plt.rcParams['axes.unicode_minus'] = False  # 解决保存图像是负号'-'显示为方块的问题,或者转换负号为字符串
groups1 = df.groupby(['sex']).get_group(('Male'))
groups01 = (groups1['total_bill']).groupby(tips['day'])
x = groups01.sum()
labels = ['Fri', 'Sat', 'Sun', 'Thur']
plt.pie(x, labels=labels, autopct='%.1f%%')
plt.title("子图4.")
plt.show()