今回は、NumPy と matplotlibライブラリで棒グラフを描いてみます。
シチュエーションとしては、あるテストの国ごとの平均点を棒グラフにしてみたいと思います。で、Excel上に、スコアのデータと国籍のデータが下方向に並んでいるとします。
| USA | 42 |
| Denmark | 42 |
| Japan | 40 |
| Denmark | 38 |
| Italy | 38 |
・
・
棒グラフ
棒グラフを描くには、matplotlib.axes.Axesクラスの bar() を使います。
bar(left, height, width=0.8, bottom=0, **kwargs) left: それぞれの棒のX座標の位置（配列で指定） height: それぞれの棒の高さ（配列で指定） width: 棒の幅 color: 棒の色 yerr: ひげの長さ
シンプルな棒グラフはこんな感じです。
from matplotlib import pyplot as plt fig = plt.figure() ax = fig.add_subplot(111) ax.bar([1, 2, 3, 4, 5], [4, 5, 6, 7, 8]) plt.show()
こんなグラフが出来上がります。
NumPyで国ごとのデータを集計
以下のように国の配列(nations) とスコアの配列(scores) が numpy.ndarray型で取得できたとします。
>>> nations
array([u'USA', u'Denmark', u'Japan', u'Denmark', u'Italy', u'Russia',
u'Turkey', u'Denmark', u'Greece', u'Japan', u'?', u'Germany',
u'Japan', u'Turkey', u'UK', u'Japan', u'Denmark', u'Turkey', u'UK',
u'Germany', u'?', u'Turkey', u'Norway', u'USA/Singapore', u'Sweden',
u'Denmark', u'Japan', u'Hong Kong', u'Finland', u'Japan', u'Italy',
u'Denmark', u'Belgium', u'Norway', u'Japan', u'Denmark', u'Norway',
u'Turkey', u'Czech', u'Germany', u'Japan', u'Japan', u'Haiti',
u'Japan', u'Denmark', u'UK', u'Germany', u'Turkey', u'Denmark',
u'Japan', u'Germany', u'Japan', u'Czech', u'UK', u'Sweden',
u'Switzerland', u'Turkey', u'Japan', u'UK', u'Japan', u'Sweden',
u'Japan', u'Denmark', u'UK', u'Japan', u'Switzerland', u'Turkey',
u'Norway', u'Sweden', u'Germany', u'France', u'Japan', u'UK',
u'Japan', u'Japan', u'Switzerland', u'Czech', u'Japan', u'Japan',
u'Denmark', u'Japan', u'Japan', u'Sweden', u'Denmark', u'Norway',
u'Israel', u'Japan', u'Japan', u'Turkey', u'Turkey', u'Turkey',
u'Turkey', u'Lebanon', u'Turkey', u'Japan', u'Turkey', u'Japan',
u'Japan', u'Japan', u'Japan', u'Japan', u'Turkey', u'Spain',
u'Belgium', u'Belgium'],
dtype='
>>> scores
array([42, 42, 40, 38, 38, 38, 38, 35, 35, 35, 35, 34, 34, 34, 34, 33, 33,
33, 33, 32, 32, 32, 30, 30, 30, 30, 30, 30, 30, 30, 29, 29, 29, 28,
28, 28, 28, 28, 27, 27, 27, 27, 27, 27, 26, 26, 26, 26, 26, 26, 25,
24, 24, 24, 24, 23, 23, 23, 23, 23, 23, 23, 23, 23, 22, 22, 22, 22,
22, 22, 22, 22, 21, 21, 20, 20, 20, 20, 20, 19, 19, 19, 19, 19, 18,
18, 17, 17, 16, 16, 16, 16, 15, 15, 15, 15, 14, 13, 13, 13, 13, 12,
12, 11, 9])
ここから、国名が 'Japan' のスコアを取り出すには、numpy.where を使います。
条件に合致する nations のインデックスを返してくれるので、scores[numpy.where(nations == 'Japan')] とすることで、スコアの配列が得られます。
>>> numpy.where(nations == 'Japan')
(array([ 2, 9, 12, 15, 26, 29, 34, 40, 41, 43, 49, 51, 57,
59, 61, 64, 71, 73, 74, 77, 78, 80, 81, 86, 87, 94,
96, 97, 98, 99, 100]),)
>>> scores[numpy.where(nations == 'Japan')]
array([40, 35, 34, 33, 30, 30, 28, 27, 27, 27, 26, 24, 23, 23, 23, 22, 22,
21, 20, 20, 20, 19, 19, 17, 17, 15, 14, 13, 13, 13, 13])
しかしもっと単純に、numpy.whereの代わりに nations == 'Japan' を使っても同じ結果を求めることができます。
>>> nations == 'Japan'
array([False, False, True, False, False, False, False, False, False,
True, False, False, True, False, False, True, False, False,
False, False, False, False, False, False, False, False, True,
False, False, True, False, False, False, False, True, False,
False, False, False, False, True, True, False, True, False,
False, False, False, False, True, False, True, False, False,
False, False, False, True, False, True, False, True, False,
False, True, False, False, False, False, False, False, True,
False, True, True, False, False, True, True, False, True,
True, False, False, False, False, True, True, False, False,
False, False, False, False, True, False, True, True, True,
True, True, False, False, False, False], dtype=bool)
>>> scores[nations == 'Japan']
array([40, 35, 34, 33, 30, 30, 28, 27, 27, 27, 26, 24, 23, 23, 23, 22, 22,
21, 20, 20, 20, 19, 19, 17, 17, 15, 14, 13, 13, 13, 13])
こうなったら、もう何でもできますね。
>>> scores[nations == 'Japan'].size 31 >>> scores[nations == 'Japan'].sum() 708 >>> scores[nations == 'Japan'].mean() 22.838709677419356 >>> scores[nations == 'Japan'].std() 6.9935305320567211
サンプル
以上を踏まえたサンプルがこちら。
test_bar_chart.py
#! /usr/bin/env python # -*- coding: utf-8 -*- import collections import numpy import xlrd from matplotlib import pyplot as plt def get_data(sheet, rowx, colx): data = [] for row in range(rowx, sheet.nrows): value = sheet.cell(row, colx).value if value != '': data.append(value) data = numpy.array(data) return data if __name__ == '__main__': book = xlrd.open_workbook('/Users/akiyoko/Documents/temp/2nd_test.xls') sheet = book.sheet_by_name('Statistics (total score)') scores = get_data(sheet, 9, 5) # データの起点はF10 nations = get_data(sheet, 9, 3) # データの起点はD10 print 'scores=%s' % scores print 'nations=%s' % nations total_size = scores.size print 'N=%d' % total_size # ラベル # labels = list(set(nations)) でもよかったが、 # 並び順がランダムなのもどうかと思ったので、度数の大きい順に並べ替えてみる counter = collections.Counter(nations) ranked_data = counter.most_common() labels = [x[0] for x in ranked_data] print 'labels=%s' % labels # 共通初期設定 plt.rc('font', **{'family': 'serif'}) # キャンバス fig = plt.figure() # ラベルが隠れてしまうのを補正 fig.subplots_adjust(bottom=0.22) # プロット領域（1x1分割の1番目に領域を配置せよという意味） ax = fig.add_subplot(111) # 棒グラフ ind = numpy.arange(len(labels)) print 'ind=%s' % ind bar_width = 0.8 mus = [] sigmas = [] for label in labels: print 'nation=%s' % label # 国ごとのスコア scores_by_nation = scores[nations == label] print 'scores_by_nation=%s' % scores_by_nation # 平均 mu = numpy.mean(scores_by_nation) mus.append(mu) print 'mean value=%.1f' % mu # 標準偏差 sigma = numpy.std(scores_by_nation) sigmas.append(sigma) print 'standard deviation=%.2f' % sigma b = ax.bar(ind, mus, bar_width, yerr=sigmas) # ラベル ax.set_xticks(ind) # ax.set_xticks(ind + bar_width / 2) ax.set_xticklabels(labels, rotation=75) # タイトル ax.set_title('Scores by Nation: N=%s' % total_size, size=16) plt.show()
実行結果
$ python test_bar_chart.py scores=[ 42. 42. 40. 38. 38. 38. 38. 35. 35. 35. 35. 34. 34. 34. 34. 33. 33. 33. 33. 32. 32. 32. 30. 30. 30. 30. 30. 30. 30. 30. 29. 29. 29. 28. 28. 28. 28. 28. 27. 27. 27. 27. 27. 27. 26. 26. 26. 26. 26. 26. 25. 24. 24. 24. 24. 23. 23. 23. 23. 23. 23. 23. 23. 23. 22. 22. 22. 22. 22. 22. 22. 22. 21. 21. 20. 20. 20. 20. 20. 19. 19. 19. 19. 19. 18. 18. 17. 17. 16. 16. 16. 16. 15. 15. 15. 15. 14. 13. 13. 13. 13. 12. 12. 11. 9.] nations=[u'USA' u'Denmark' u'Japan' u'Denmark' u'Italy' u'Russia' u'Turkey' u'Denmark' u'Greece' u'Japan' u'?' u'Germany' u'Japan' u'Turkey' u'UK' u'Japan' u'Denmark' u'Turkey' u'UK' u'Germany' u'?' u'Turkey' u'Norway' u'USA/Singapore' u'Sweden' u'Denmark' u'Japan' u'Hong Kong' u'Finland' u'Japan' u'Italy' u'Denmark' u'Belgium' u'Norway' u'Japan' u'Denmark' u'Norway' u'Turkey' u'Czech' u'Germany' u'Japan' u'Japan' u'Haiti' u'Japan' u'Denmark' u'UK' u'Germany' u'Turkey' u'Denmark' u'Japan' u'Germany' u'Japan' u'Czech' u'UK' u'Sweden' u'Switzerland' u'Turkey' u'Japan' u'UK' u'Japan' u'Sweden' u'Japan' u'Denmark' u'UK' u'Japan' u'Switzerland' u'Turkey' u'Norway' u'Sweden' u'Germany' u'France' u'Japan' u'UK' u'Japan' u'Japan' u'Switzerland' u'Czech' u'Japan' u'Japan' u'Denmark' u'Japan' u'Japan' u'Sweden' u'Denmark' u'Norway' u'Israel' u'Japan' u'Japan' u'Turkey' u'Turkey' u'Turkey' u'Turkey' u'Lebanon' u'Turkey' u'Japan' u'Turkey' u'Japan' u'Japan' u'Japan' u'Japan' u'Japan' u'Turkey' u'Spain' u'Belgium' u'Belgium'] N=105 labels=[u'Japan', u'Turkey', u'Denmark', u'UK', u'Germany', u'Norway', u'Sweden', u'Belgium', u'Switzerland', u'Czech', u'Italy', u'?', u'USA', u'France', u'Israel', u'Haiti', u'Hong Kong', u'USA/Singapore', u'Finland', u'Russia', u'Lebanon', u'Spain', u'Greece'] ind=[ 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22] nation=Japan scores_by_nation=[ 40. 35. 34. 33. 30. 30. 28. 27. 27. 27. 26. 24. 23. 23. 23. 22. 22. 21. 20. 20. 20. 19. 19. 17. 17. 15. 14. 13. 13. 13. 13.] mean value=22.8 standard deviation=6.99 nation=Turkey scores_by_nation=[ 38. 34. 33. 32. 28. 26. 23. 22. 16. 16. 16. 16. 15. 15. 12.] mean value=22.8 standard deviation=8.19 nation=Denmark scores_by_nation=[ 42. 38. 35. 33. 30. 29. 28. 26. 26. 23. 19. 19.] mean value=29.0 standard deviation=6.82 nation=UK scores_by_nation=[ 34. 33. 26. 24. 23. 23. 21.] mean value=26.3 standard deviation=4.77 nation=Germany scores_by_nation=[ 34. 32. 27. 26. 25. 22.] mean value=27.7 standard deviation=4.11 nation=Norway scores_by_nation=[ 30. 28. 28. 22. 18.] mean value=25.2 standard deviation=4.49 nation=Sweden scores_by_nation=[ 30. 24. 23. 22. 19.] mean value=23.6 standard deviation=3.61 nation=Belgium scores_by_nation=[ 29. 11. 9.] mean value=16.3 standard deviation=8.99 nation=Switzerland scores_by_nation=[ 23. 22. 20.] mean value=21.7 standard deviation=1.25 nation=Czech scores_by_nation=[ 27. 24. 20.] mean value=23.7 standard deviation=2.87 nation=Italy scores_by_nation=[ 38. 29.] mean value=33.5 standard deviation=4.50 nation=? scores_by_nation=[ 35. 32.] mean value=33.5 standard deviation=1.50 nation=USA scores_by_nation=[ 42.] mean value=42.0 standard deviation=0.00 nation=France scores_by_nation=[ 22.] mean value=22.0 standard deviation=0.00 nation=Israel scores_by_nation=[ 18.] mean value=18.0 standard deviation=0.00 nation=Haiti scores_by_nation=[ 27.] mean value=27.0 standard deviation=0.00 nation=Hong Kong scores_by_nation=[ 30.] mean value=30.0 standard deviation=0.00 nation=USA/Singapore scores_by_nation=[ 30.] mean value=30.0 standard deviation=0.00 nation=Finland scores_by_nation=[ 30.] mean value=30.0 standard deviation=0.00 nation=Russia scores_by_nation=[ 38.] mean value=38.0 standard deviation=0.00 nation=Lebanon scores_by_nation=[ 15.] mean value=15.0 standard deviation=0.00 nation=Spain scores_by_nation=[ 12.] mean value=12.0 standard deviation=0.00 nation=Greece scores_by_nation=[ 35.] mean value=35.0 standard deviation=0.00
ラベルが少し隠れているのが残念ですね。今後、改良していきます。
ラベルが隠れるのを修正できました。
# ラベルが隠れてしまうのを補正 fig.subplots_adjust(bottom=0.22)
とやればよいのでした。bettamodokiのメモ が参考になりました。（2013/6/25追記）
ちなみに、修正前のグラフはこちら。
