如何制作 Pandas DataFrame 的简单等高线图，其中数字单元格值为 Z 并标记为 rows/columns 作为 X 和 Y 坐标？
How do I make a make a simple contour chart of a Pandas DataFrame with numeric cell values as Z and labeled rows/columns as X and Y coordinates?
我有一个 Pandas DataFrame，luminance_df，它看起来像这样：
barelyvisible
ultralight
light
abitlight
medium
abitdark
dark
evendarker
ultradark
almostblack
orange
96
92
83
72
61
53
48
40
34
28
gold
96
89
77
65
56
50
44
37
31
26
yellow
95
88
77
64
53
47
40
33
29
26
chartreuse
95
89
80
67
55
44
35
27
23
20
green
97
93
85
73
58
45
36
29
24
20
forest
96
90
80
67
52
39
30
24
20
16
aqua
97
89
78
64
50
40
32
26
22
19
teal
96
90
82
69
53
43
36
31
27
24
lightblue
97
94
86
74
60
48
39
32
27
24
blue
97
93
87
78
68
60
53
48
40
33
indigo
97
94
89
82
74
67
59
51
41
34
purple
98
95
92
85
76
66
58
50
42
35
royalpurple
98
95
92
85
75
65
56
47
39
32
magenta
98
95
91
83
73
61
49
40
33
28
pink
97
95
90
82
70
60
51
42
35
30
dustypink
97
95
90
82
71
60
50
41
35
30
red
97
94
89
82
71
60
51
42
35
31
到目前为止，我正在构建一个多图表 HTML 文件，如下所示：
with open(os.path.join(cwd, 'testout.html'), 'w') as outfile:
outfile.write("<p> </p><hr/><p> </p>".join(['<h1>Colors</h1>'+hex_styler.to_html(), '<h1>Hue</h1>'+hue_styler.to_html(), '<h1>Saturation</h1>'+saturation_styler.to_html(
), '<h1>Luminance</h1>'+luminance_styler.to_html(), '<h1>Perceived Brightness</h1>'+perceived_brightness_pivot_styler.to_html(), '<h1>Base Data</h1>'+basic_df.to_html()]))
我想在 luminance_styler.to_html() 之后显示一个 elevation/contour 样式的亮度图，很像我在 Excel:
中制作的这张图
我希望颜色保持“从上到下”排序，作为 y 轴上的值，而深色保持“从左到右”排序，作为 x 轴上的值，就像在示例中一样以上。
问题
我不是数据科学家，也不经常使用 Python。我为自己首先制作 luminance_df 感到自豪，但我并没有想出如何简单地制作 Python ...两个方向的标签都是字符串的DataFrame ...作为z轴并制作它的等高线图。
我 Google 所做的一切都会导致非常复杂的数据科学细微问题。
有人可以通过给我基本的“hello world”代码让我走上正轨，让我至少了解 Python 中 luminance_df 的数据我在 Excel?
中使用了“插入图表”按钮
如果你能找到我，那么我就有了一个 image_base64 = base64.b64encode(img.read()).decode("utf-8")-able 的 img = BytesIO()，我可以 f'<img src="data:image/png;base64, {image_base64}" />' 自己将它变成 testout.html 的字符串连接.
我在 Windows 并且我自己设置为能够 pip install。
注释
公平地说，我发现这些等高线图比 Excel 制作的更有吸引力，也更容易阅读，但我可以接受一些“野蛮”的东西-看起来像 Excel 版本，只要它使“上升”和“下降”变得明显，并且只要它使用 ROYIGBV 彩虹来指示“更少”与“更多”（讨厌关于默认 Excel 颜色的我的意见 -- 是的，我知道，这可能是一个可访问性问题）:
虽然我希望图表的颜色遵循某种“彩虹”（因为我个人认为它们易于阅读），但任何“彩虹”图表上的阴影”应该完全忽略 y 轴的标签恰好描述颜色的事实。没有任何相关性。我只是在绘制 16 到 98 之间的数字事实；图表的颜色应该仅表示这两个极端之间“高度”的变化。
到目前为止的努力
到目前为止我发现的唯一一个看起来相似的“简单”问题是 Convert pandas DataFrame to a 3d graph using Index and Columns as X,Y and values as Z?，但这段代码对我来说根本不起作用，所以我什至不知道它输出什么，视觉上，所以我不知道它是否相关：
from mpl_toolkits.mplot3d import Axes3D
import matplotlib.pyplot as plt
lumX = luminance_df.columns
lumY = luminance_df.index
lumZ = luminance_df.values
fig = plt.figure()
ax = plt.axes(projection = '3d')
ax.contour3D(lumX,lumY,lumZ)
我的脚本出错并显示一条消息：ValueError: could not convert string to float: 'orange'，而且我不知道我在做什么以适应这个答案似乎是围绕数字 X 的假设编写的- 和 Y 轴键。 （此外，它可能不会生成我希望的图表类型——正如我所说，无法判断，因为它甚至不执行，而且答案中没有视觉样本。）
数据集
准备 pandas.DataFrame():
{"barelyvisible":{"orange":96,"gold":96,"yellow":95,"chartreuse":95,"green":97,"forest":96,"aqua":97,"teal":96,"lightblue":97,"blue":97,"indigo":97,"purple":98,"royalpurple":98,"magenta":98,"pink":97,"dustypink":97,"red":97},"ultralight":{"orange":92,"gold":89,"yellow":88,"chartreuse":89,"green":93,"forest":90,"aqua":89,"teal":90,"lightblue":94,"blue":93,"indigo":94,"purple":95,"royalpurple":95,"magenta":95,"pink":95,"dustypink":95,"red":94},"light":{"orange":83,"gold":77,"yellow":77,"chartreuse":80,"green":85,"forest":80,"aqua":78,"teal":82,"lightblue":86,"blue":87,"indigo":89,"purple":92,"royalpurple":92,"magenta":91,"pink":90,"dustypink":90,"red":89},"abitlight":{"orange":72,"gold":65,"yellow":64,"chartreuse":67,"green":73,"forest":67,"aqua":64,"teal":69,"lightblue":74,"blue":78,"indigo":82,"purple":85,"royalpurple":85,"magenta":83,"pink":82,"dustypink":82,"red":82},"medium":{"orange":61,"gold":56,"yellow":53,"chartreuse":55,"green":58,"forest":52,"aqua":50,"teal":53,"lightblue":60,"blue":68,"indigo":74,"purple":76,"royalpurple":75,"magenta":73,"pink":70,"dustypink":71,"red":71},"abitdark":{"orange":53,"gold":50,"yellow":47,"chartreuse":44,"green":45,"forest":39,"aqua":40,"teal":43,"lightblue":48,"blue":60,"indigo":67,"purple":66,"royalpurple":65,"magenta":61,"pink":60,"dustypink":60,"red":60},"dark":{"orange":48,"gold":44,"yellow":40,"chartreuse":35,"green":36,"forest":30,"aqua":32,"teal":36,"lightblue":39,"blue":53,"indigo":59,"purple":58,"royalpurple":56,"magenta":49,"pink":51,"dustypink":50,"red":51},"evendarker":{"orange":40,"gold":37,"yellow":33,"chartreuse":27,"green":29,"forest":24,"aqua":26,"teal":31,"lightblue":32,"blue":48,"indigo":51,"purple":50,"royalpurple":47,"magenta":40,"pink":42,"dustypink":41,"red":42},"ultradark":{"orange":34,"gold":31,"yellow":29,"chartreuse":23,"green":24,"forest":20,"aqua":22,"teal":27,"lightblue":27,"blue":40,"indigo":41,"purple":42,"royalpurple":39,"magenta":33,"pink":35,"dustypink":35,"red":35},"almostblack":{"orange":28,"gold":26,"yellow":26,"chartreuse":20,"green":20,"forest":16,"aqua":19,"teal":24,"lightblue":24,"blue":33,"indigo":34,"purple":35,"royalpurple":32,"magenta":28,"pink":30,"dustypink":30,"red":31}}
我相信你只需要做一个countourf:
plt.contourf(df, cmap='RdYlBu')
plt.xticks(range(df.shape[1]), df.columns, rotation=90)
plt.yticks(range(df.shape[0]), df.index)
plt.show()
输出：
或热图：
import seaborn as sns
sns.heatmap(df, cmap='RdYlBu')
输出：
我有一个 Pandas DataFrame，luminance_df，它看起来像这样：
| barelyvisible | ultralight | light | abitlight | medium | abitdark | dark | evendarker | ultradark | almostblack | |
|---|---|---|---|---|---|---|---|---|---|---|
| orange | 96 | 92 | 83 | 72 | 61 | 53 | 48 | 40 | 34 | 28 |
| gold | 96 | 89 | 77 | 65 | 56 | 50 | 44 | 37 | 31 | 26 |
| yellow | 95 | 88 | 77 | 64 | 53 | 47 | 40 | 33 | 29 | 26 |
| chartreuse | 95 | 89 | 80 | 67 | 55 | 44 | 35 | 27 | 23 | 20 |
| green | 97 | 93 | 85 | 73 | 58 | 45 | 36 | 29 | 24 | 20 |
| forest | 96 | 90 | 80 | 67 | 52 | 39 | 30 | 24 | 20 | 16 |
| aqua | 97 | 89 | 78 | 64 | 50 | 40 | 32 | 26 | 22 | 19 |
| teal | 96 | 90 | 82 | 69 | 53 | 43 | 36 | 31 | 27 | 24 |
| lightblue | 97 | 94 | 86 | 74 | 60 | 48 | 39 | 32 | 27 | 24 |
| blue | 97 | 93 | 87 | 78 | 68 | 60 | 53 | 48 | 40 | 33 |
| indigo | 97 | 94 | 89 | 82 | 74 | 67 | 59 | 51 | 41 | 34 |
| purple | 98 | 95 | 92 | 85 | 76 | 66 | 58 | 50 | 42 | 35 |
| royalpurple | 98 | 95 | 92 | 85 | 75 | 65 | 56 | 47 | 39 | 32 |
| magenta | 98 | 95 | 91 | 83 | 73 | 61 | 49 | 40 | 33 | 28 |
| pink | 97 | 95 | 90 | 82 | 70 | 60 | 51 | 42 | 35 | 30 |
| dustypink | 97 | 95 | 90 | 82 | 71 | 60 | 50 | 41 | 35 | 30 |
| red | 97 | 94 | 89 | 82 | 71 | 60 | 51 | 42 | 35 | 31 |
到目前为止，我正在构建一个多图表 HTML 文件，如下所示：
with open(os.path.join(cwd, 'testout.html'), 'w') as outfile:
outfile.write("<p> </p><hr/><p> </p>".join(['<h1>Colors</h1>'+hex_styler.to_html(), '<h1>Hue</h1>'+hue_styler.to_html(), '<h1>Saturation</h1>'+saturation_styler.to_html(
), '<h1>Luminance</h1>'+luminance_styler.to_html(), '<h1>Perceived Brightness</h1>'+perceived_brightness_pivot_styler.to_html(), '<h1>Base Data</h1>'+basic_df.to_html()]))
我想在 luminance_styler.to_html() 之后显示一个 elevation/contour 样式的亮度图，很像我在 Excel:
我希望颜色保持“从上到下”排序，作为 y 轴上的值，而深色保持“从左到右”排序，作为 x 轴上的值，就像在示例中一样以上。
问题
我不是数据科学家，也不经常使用 Python。我为自己首先制作 luminance_df 感到自豪，但我并没有想出如何简单地制作 Python ...两个方向的标签都是字符串的DataFrame ...作为z轴并制作它的等高线图。
我 Google 所做的一切都会导致非常复杂的数据科学细微问题。
有人可以通过给我基本的“hello world”代码让我走上正轨，让我至少了解 Python 中 luminance_df 的数据我在 Excel?
如果你能找到我，那么我就有了一个 image_base64 = base64.b64encode(img.read()).decode("utf-8")-able 的 img = BytesIO()，我可以 f'<img src="data:image/png;base64, {image_base64}" />' 自己将它变成 testout.html 的字符串连接.
我在 Windows 并且我自己设置为能够 pip install。
注释
公平地说，我发现这些等高线图比 Excel 制作的更有吸引力，也更容易阅读，但我可以接受一些“野蛮”的东西-看起来像 Excel 版本，只要它使“上升”和“下降”变得明显，并且只要它使用 ROYIGBV 彩虹来指示“更少”与“更多”（讨厌关于默认 Excel 颜色的我的意见 -- 是的，我知道，这可能是一个可访问性问题）:
虽然我希望图表的颜色遵循某种“彩虹”（因为我个人认为它们易于阅读），但任何“彩虹”图表上的阴影”应该完全忽略 y 轴的标签恰好描述颜色的事实。没有任何相关性。我只是在绘制 16 到 98 之间的数字事实；图表的颜色应该仅表示这两个极端之间“高度”的变化。
到目前为止的努力
到目前为止我发现的唯一一个看起来相似的“简单”问题是 Convert pandas DataFrame to a 3d graph using Index and Columns as X,Y and values as Z?，但这段代码对我来说根本不起作用，所以我什至不知道它输出什么，视觉上，所以我不知道它是否相关：
from mpl_toolkits.mplot3d import Axes3D
import matplotlib.pyplot as plt
lumX = luminance_df.columns
lumY = luminance_df.index
lumZ = luminance_df.values
fig = plt.figure()
ax = plt.axes(projection = '3d')
ax.contour3D(lumX,lumY,lumZ)
我的脚本出错并显示一条消息：ValueError: could not convert string to float: 'orange'，而且我不知道我在做什么以适应这个答案似乎是围绕数字 X 的假设编写的- 和 Y 轴键。 （此外，它可能不会生成我希望的图表类型——正如我所说，无法判断，因为它甚至不执行，而且答案中没有视觉样本。）
数据集
准备 pandas.DataFrame():
{"barelyvisible":{"orange":96,"gold":96,"yellow":95,"chartreuse":95,"green":97,"forest":96,"aqua":97,"teal":96,"lightblue":97,"blue":97,"indigo":97,"purple":98,"royalpurple":98,"magenta":98,"pink":97,"dustypink":97,"red":97},"ultralight":{"orange":92,"gold":89,"yellow":88,"chartreuse":89,"green":93,"forest":90,"aqua":89,"teal":90,"lightblue":94,"blue":93,"indigo":94,"purple":95,"royalpurple":95,"magenta":95,"pink":95,"dustypink":95,"red":94},"light":{"orange":83,"gold":77,"yellow":77,"chartreuse":80,"green":85,"forest":80,"aqua":78,"teal":82,"lightblue":86,"blue":87,"indigo":89,"purple":92,"royalpurple":92,"magenta":91,"pink":90,"dustypink":90,"red":89},"abitlight":{"orange":72,"gold":65,"yellow":64,"chartreuse":67,"green":73,"forest":67,"aqua":64,"teal":69,"lightblue":74,"blue":78,"indigo":82,"purple":85,"royalpurple":85,"magenta":83,"pink":82,"dustypink":82,"red":82},"medium":{"orange":61,"gold":56,"yellow":53,"chartreuse":55,"green":58,"forest":52,"aqua":50,"teal":53,"lightblue":60,"blue":68,"indigo":74,"purple":76,"royalpurple":75,"magenta":73,"pink":70,"dustypink":71,"red":71},"abitdark":{"orange":53,"gold":50,"yellow":47,"chartreuse":44,"green":45,"forest":39,"aqua":40,"teal":43,"lightblue":48,"blue":60,"indigo":67,"purple":66,"royalpurple":65,"magenta":61,"pink":60,"dustypink":60,"red":60},"dark":{"orange":48,"gold":44,"yellow":40,"chartreuse":35,"green":36,"forest":30,"aqua":32,"teal":36,"lightblue":39,"blue":53,"indigo":59,"purple":58,"royalpurple":56,"magenta":49,"pink":51,"dustypink":50,"red":51},"evendarker":{"orange":40,"gold":37,"yellow":33,"chartreuse":27,"green":29,"forest":24,"aqua":26,"teal":31,"lightblue":32,"blue":48,"indigo":51,"purple":50,"royalpurple":47,"magenta":40,"pink":42,"dustypink":41,"red":42},"ultradark":{"orange":34,"gold":31,"yellow":29,"chartreuse":23,"green":24,"forest":20,"aqua":22,"teal":27,"lightblue":27,"blue":40,"indigo":41,"purple":42,"royalpurple":39,"magenta":33,"pink":35,"dustypink":35,"red":35},"almostblack":{"orange":28,"gold":26,"yellow":26,"chartreuse":20,"green":20,"forest":16,"aqua":19,"teal":24,"lightblue":24,"blue":33,"indigo":34,"purple":35,"royalpurple":32,"magenta":28,"pink":30,"dustypink":30,"red":31}}
我相信你只需要做一个countourf:
plt.contourf(df, cmap='RdYlBu')
plt.xticks(range(df.shape[1]), df.columns, rotation=90)
plt.yticks(range(df.shape[0]), df.index)
plt.show()
输出：
或热图：
import seaborn as sns
sns.heatmap(df, cmap='RdYlBu')
输出：