Python 3.5.2 中的本福德定律绘图

Question

我被分配的一项任务难住了，该任务要求我输出从 CSV 文件中读取的数据集...

...根据本福德定律，得到如下条形图：

Benford 条形图示例

这是我目前得到的代码：

import matplotlib.pyplot as plt
import math
import csv
import locale

with open("immigrants.csv", newline='') as csvfile:
    immidata = csv.reader(csvfile)
    X_labels = []
    Y = []
    for row in immidata:
        X_labels.append(row[0])
        Y.append(locale.atoi(row[1]))

numbers = [float(n) for n in range(1, 10)]
benford = [math.log10(1 + 1 / d) for d in numbers]
plt.plot(numbers, benford, 'ro', label = "Benford's Law")
plt.bar(numbers, range(1, 11), align = 'left', normed = True, 
    rwidth = 0.7, label = "Actual data")
plt.bar(benford, range(1, 11), align = 'left', normed = True, 
    rwidth = 0.7, label = "Predicted data")
plt.title("Immigrants in countries")
plt.xlabel("Digit")
plt.ylabel("Probability")
plt.grid(True)
plt.xlim(0, 10)
plt.xticks(numbers)
plt.legend()
plt.show()

以下是 CSV 文件中的一些信息，其中显示了每个国家/地区的移民人数（分别为国家、移民人数、世界移民总数的百分比和移民占全国人口的百分比）：

United States,"45,785,090",19.8,14.3
Russia,"11,048,064",4.8,7.7
Germany,"9,845,244",4.3,11.9
Saudi Arabia,"9,060,433",3.9,31.4
United Arab Emirates,"7,826,981",3.4,83.7
United Kingdom,"7,824,131",3.4,12.4

我现在的输出：

line 19, in <module>
  Y.append(locale.atoi(row[1]))
line 321, in atoi
  return int(delocalize(string))
ValueError: invalid literal for int() with base 10: 'Number of
immigrants'

Process finished with exit code 1

我对此还比较陌生，因此非常感谢任何有助于我获得输出的建议！

谢谢！

Output

需要看起来像样本的输出。

Answer 1

这是一个使用 pandas 的解决方案：

这是本福德定律与移民人数的对比

编辑： 您的文件的标题行可能由 num_immigrants 列中的字符串“移民人数”指示。删除读取数据行中的header=None 选项。

import pandas as pd
import numpy as np
import matplotlib.pyplot as plt

# set the width of the bars, you're gonna have to massage this
width = 0.35

immi = pd.read_csv('immigrants.csv')

# name columns
immi.columns = ['country', 'num_immigrants', 'perc_world', 'perc_nat_pop']

# convert num_immigrants to float
immi.num_immigrants=  immi.num_immigrants.str.replace(',', '').apply(float)
total = immi.num_immigrants.sum()

# scale the immigration to between 0 and 1
immi['immi_scaled'] = immi['num_immigrants'].apply(lambda x: x/total)

indx = np.arange(1, len(immi) + 1)
benford = [np.log10(1 + (1.0 / d)) for d in indx]

plt.bar(indx, benford, width, color='r', label="Benford's Law")
plt.bar(np.arange(1, immi.shape[0]+1)+ width, 
                immi.immi_scaled, width, color='b', label="Predicted data")
# center the xtick labels
ax = plt.gca()
ax.set_xticks(indx + width / 2)
ax.set_xticklabels((indx))

# limit the  number of bars if you have more data
plt.xlim(1, 9)
plt.title("Immigrants in countries")
plt.ylabel("Probability")
plt.grid(True)
plt.legend()
plt.show()

Answer 2

读取数据：

import locale

with open("immigrants.csv", newline='') as csvfile:
    immidata = csv.reader(csvfile)  # defaults are fine!
    X_labels = []
    Y = []
    for row in immidata:
        X_labels.append(row[0])
        Y.append(locale.atoi(row[1]))

为您提供 X_labels 和 Y（转换为 int）。
注意：不需要close() with 块会自动执行此操作。

祝你好运。顺便说一句：digits 在您共享的代码中未定义 - 您应该尽一切努力使其成为 MCVE