1

from scipy.stats import sem

2

import pandas as pd

3

import matplotlib.pyplot as plt

4

# Import Data

5

df_raw = pd.read_csv('F:\数据杂坛\datasets\orders_45d.csv',

6

parse_dates=['purchase_time', 'purchase_date'])

7

8

# Prepare Data: Daily Mean and SE Bands

9

df_mean = df_raw.groupby('purchase_date').quantity.mean()

10

df_se = df_raw.groupby('purchase_date').quantity.apply(sem).mul(1.96)

11

12

# Plot

13

plt.figure(figsize=(10, 6), dpi=80)

14

plt.ylabel("Daily Orders", fontsize=12)

15

x = [d.date().strftime('%Y-%m-%d') for d in df_mean.index]

16

plt.plot(x, df_mean, color="#c72e29", lw=2)

17

plt.fill_between(x, df_mean - df_se, df_mean + df_se, color="#f8f2e4")

18

19

# Decorations

20

# Lighten borders

21

plt.gca().spines["top"].set_alpha(0)

22

plt.gca().spines["bottom"].set_alpha(1)

23

plt.gca().spines["right"].set_alpha(0)

24

plt.gca().spines["left"].set_alpha(1)

25

plt.xticks(x[::6], [str(d) for d in x[::6]], fontsize=12)

26

plt.title("Daily Order Quantity of Brazilian Retail with Error Bands (95% confidence)",fontsize=14)

27

28

# Axis limits

29

s, e = plt.gca().get_xlim()

30

plt.xlim(s, e - 2)

31

plt.ylim(4, 10)

32

33

# Draw Horizontal Tick lines

34

for y in range(5, 10, 1):

35

plt.hlines(y,

36

xmin=s,

37

xmax=e,

38

colors='black',

39

alpha=0.5,

40

linestyles="--",

41

lw=0.5)

42

43

plt.show()