Plotly：如何在指定阈值的折线图上显示不同的颜色段？答案

【问题标题】：Plotly: How to display different color segments on a line chart for specified thresholds?Plotly：如何在指定阈值的折线图上显示不同的颜色段？
【发布时间】：2021-02-21 21:27:56
【问题描述】：

我有一个显示随时间增加的百分比的多线图。我想在我的代码中设置一个阈值以具有上限和下限。如果线超出这些范围，我希望线图的特定部分与其父级颜色不同。

这就是我正在做的事情：

import plotly.express as px
import plotly.graph_objects as go



fig = px.line(df14, x = "Date", y = "Percent", color = "id", 
          title = "id Growth in Percentage (US)", 
          labels = {"Percent": "Percent Growth"})

fig.update_layout(
font_family="Arial",
font_color="black",
title_font_family="Arial",
title_font_color="black",
legend_title_font_color="black"                                           #style the text (legend, title etc)
)


fig.update_xaxes(title_font_family="Arial")                               #style ance center title
fig.update_layout(
title={
    'text': "id Growth Percentage in US (Line Graph)",
    'y':0.9,
    'x':0.5,
    'xanchor': 'center',
    'yanchor': 'top'})


fig.update_traces(mode='markers+lines')                                    #add dots to line

fig.show()

这是视觉效果：

让我放大一行以便更好地解释：

我想为每个 id 设置一个阈值，并且如果线条高于或低于此阈值，那么线条的那部分颜色会有所不同 .例如：

id IAA 的上限可能是 5，IAA 的下限可能是 0。 任何大于 5 或小于 0 的值都会以特定颜色突出显示。

id SSS 的上限可能是 10，SSS 的下限可能是 3 任何大于 10 或小于 3 的值都应以特定颜色突出显示。

我希望每个 id 都有阈值

请看下面：

线条中突出显示的黄色部分反映了折线图超出或降低了设定阈值的位置。使用 Plotly 可以做到这一点吗？

以下是原始数据示例： 更新：

    id       Start      End         Diff        Percent     Date        
    IAA     4/1/2019    5/1/2019    160.4279    11.10809    04-01-2019 to 05-01-2019
    IAA 5/1/2019    6/1/2019    136.0248    8.476798    05-01-2019 to 06-01-2019
    IAA     6/1/2019    7/1/2019    174.0513    9.998946    06-01-2019 to 07-01-2019
    IAA     7/1/2019    8/1/2019    112.0424    5.851551    07-01-2019 to 08-01-2019
    IAA     8/1/2019    9/1/2019    141.8488    6.998691    08-01-2019 to 09-01-2019
    IAA     9/1/2019    10/1/2019   103.5522    4.774984    09-01-2019 to 10-01-2019
    IAA     10/1/2019   11/1/2019   125.6087    5.528085    10-01-2019 to 11-01-2019
    IAA     11/1/2019   12/1/2019   145.2591    6.058016    11-01-2019 to 12-01-2019
    IAA     12/1/2019   1/1/2020    115.5121    4.542251    12-01-2019 to 01-01-2020
    IAA     1/1/2020    2/1/2020    185.7191    6.985673    01-01-2020 to 02-01-2020
    IAA     2/1/2020    3/1/2020    126.7386    4.455896    02-01-2020 to 03-01-2020
    IAA     3/1/2020    4/1/2020    231.3461    7.786734    03-01-2020 to 04-01-2020
    IAA     4/1/2020    5/1/2020    97.02587      3.02981   04-01-2020 to 05-01-2020
    IAA     5/1/2020    6/1/2020    42.85235      1.298792  05-01-2020 to 06-01-2020
    IAA     6/1/2020    7/1/2020    124.666    3.729997     06-01-2020 to 07-01-2020
    IAA     7/1/2020    8/1/2020    357.9974    10.32609    07-01-2020 to 08-01-2020
    IAA     8/1/2020    9/1/2020    490.9587      12.8358   08-01-2020 to 09-01-2020
    IAA     9/1/2020    10/1/2020   204.5478    4.739428    09-01-2020 to 10-01-2020
    IAA     10/1/2020   11/1/2020   287.6025    6.362292    10-01-2020 to 11-01-2020
   SSStest  4/1/2019    5/1/2019    12.38486    5.780551    04-01-2019 to 05-01-2019
   SSStest  5/1/2019    6/1/2019    -2.61735    -1.15487    05-01-2019 to 06-01-2019
   SSStest  6/1/2019    7/1/2019    -5.6187    -2.50814     06-01-2019 to 07-01-2019
   SSStest  7/1/2019    8/1/2019    3.204252    1.467153    07-01-2019 to 08-01-2019
   SSStest  8/1/2019    9/1/2019    -25.3782    -11.4521    08-01-2019 to 09-01-2019
   SSStest  9/1/2019    10/1/2019   -10.9717    -5.59137    09-01-2019 to 10-01-2019

欢迎提出任何建议

更新

有没有办法让整行显示标记点？我试过这个：mode = 'markers+lines' 但没有得到想要的结果：

**Update, I have figured this out:** 
fig.update_traces(mode='markers+lines')

更新问题：

另外，有没有办法在这里的悬停注释上添加日期和百分比标题？我正在研究 Plotly 文档。

[![在此处输入图片描述][4]][4]

【问题讨论】：

我想看看当多个折线图超过或低于某个阈值时，Plotly 是否能够为它们显示不同的颜色。我已经研究了一段时间，并没有提出解决方案。任何输入都会有所帮助。谢谢@vestland

标签： python pandas plotly threshold

【解决方案1】：

我已经提出了一个建议，应该完全符合您的要求。下图由下面的代码示例生成。该代码使用一个字典，其中包含不同 POD 的不同上限和下限以及为不同 pod 设置不同颜色的可能性：

lim = {'IAD': {'lower': 90,'upper': 350, 'color':'yellow'},
       'SJCtest': {'lower': 10,'upper': 12, 'color':'green'}}

情节

我很想先过滤您的数据框，然后在此基础上向图中添加新的跟踪。但我的解决方案实际上会遍历每个跟踪的每个数据点，并根据上面的字典为这些点着色。所以不需要任何数据处理。看一看，运行一些测试，然后告诉我它对你的效果。

完整代码

import pandas as pd
import plotly.express as px
import plotly.graph_objects as go

df = pd.DataFrame({'POD': {0: 'IAD',
                          1: 'IAD',
                          2: 'IAD',
                          3: 'IAD',
                          4: 'IAD',
                          5: 'IAD',
                          6: 'IAD',
                          7: 'IAD',
                          8: 'IAD',
                          9: 'IAD',
                          10: 'IAD',
                          11: 'IAD',
                          12: 'IAD',
                          13: 'IAD',
                          14: 'IAD',
                          15: 'IAD',
                          16: 'IAD',
                          17: 'IAD',
                          18: 'IAD',
                          19: 'SJCtest',
                          20: 'SJCtest',
                          21: 'SJCtest',
                          22: 'SJCtest',
                          23: 'SJCtest',
                          24: 'SJCtest'},
                         'Start': {0: '4/1/2019',
                          1: '5/1/2019',
                          2: '6/1/2019',
                          3: '7/1/2019',
                          4: '8/1/2019',
                          5: '9/1/2019',
                          6: '10/1/2019',
                          7: '11/1/2019',
                          8: '12/1/2019',
                          9: '1/1/2020',
                          10: '2/1/2020',
                          11: '3/1/2020',
                          12: '4/1/2020',
                          13: '5/1/2020',
                          14: '6/1/2020',
                          15: '7/1/2020',
                          16: '8/1/2020',
                          17: '9/1/2020',
                          18: '10/1/2020',
                          19: '4/1/2019',
                          20: '5/1/2019',
                          21: '6/1/2019',
                          22: '7/1/2019',
                          23: '8/1/2019',
                          24: '9/1/2019'},
                         'End': {0: '5/1/2019',
                          1: '6/1/2019',
                          2: '7/1/2019',
                          3: '8/1/2019',
                          4: '9/1/2019',
                          5: '10/1/2019',
                          6: '11/1/2019',
                          7: '12/1/2019',
                          8: '1/1/2020',
                          9: '2/1/2020',
                          10: '3/1/2020',
                          11: '4/1/2020',
                          12: '5/1/2020',
                          13: '6/1/2020',
                          14: '7/1/2020',
                          15: '8/1/2020',
                          16: '9/1/2020',
                          17: '10/1/2020',
                          18: '11/1/2020',
                          19: '5/1/2019',
                          20: '6/1/2019',
                          21: '7/1/2019',
                          22: '8/1/2019',
                          23: '9/1/2019',
                          24: '10/1/2019'},
                         'Diff': {0: 160.4279,
                          1: 136.0248,
                          2: 174.0513,
                          3: 112.0424,
                          4: 141.8488,
                          5: 103.5522,
                          6: 125.6087,
                          7: 145.2591,
                          8: 115.5121,
                          9: 185.7191,
                          10: 126.7386,
                          11: 231.3461,
                          12: 97.02587,
                          13: 42.85235,
                          14: 124.666,
                          15: 357.9974,
                          16: 490.9587,
                          17: 204.5478,
                          18: 287.6025,
                          19: 12.38486,
                          20: -2.61735,
                          21: -5.6187,
                          22: 3.204252,
                          23: -25.3782,
                          24: -10.9717},
                         'Percent': {0: 11.108089999999999,
                          1: 8.476797999999999,
                          2: 9.998946,
                          3: 5.851551000000001,
                          4: 6.998691,
                          5: 4.774984,
                          6: 5.528085,
                          7: 6.058016,
                          8: 4.542251,
                          9: 6.985672999999999,
                          10: 4.455896,
                          11: 7.786734,
                          12: 3.02981,
                          13: 1.298792,
                          14: 3.729997,
                          15: 10.326089999999999,
                          16: 12.8358,
                          17: 4.739428,
                          18: 6.362292,
                          19: 5.780551,
                          20: -1.15487,
                          21: -2.50814,
                          22: 1.4671530000000002,
                          23: -11.4521,
                          24: -5.5913699999999995},
                         'Date': {0: '04-01-2019 to 05-01-2019',
                          1: '05-01-2019 to 06-01-2019',
                          2: '06-01-2019 to 07-01-2019',
                          3: '07-01-2019 to 08-01-2019',
                          4: '08-01-2019 to 09-01-2019',
                          5: '09-01-2019 to 10-01-2019',
                          6: '10-01-2019 to 11-01-2019',
                          7: '11-01-2019 to 12-01-2019',
                          8: '12-01-2019 to 01-01-2020',
                          9: '01-01-2020 to 02-01-2020',
                          10: '02-01-2020 to 03-01-2020',
                          11: '03-01-2020 to 04-01-2020',
                          12: '04-01-2020 to 05-01-2020',
                          13: '05-01-2020 to 06-01-2020',
                          14: '06-01-2020 to 07-01-2020',
                          15: '07-01-2020 to 08-01-2020',
                          16: '08-01-2020 to 09-01-2020',
                          17: '09-01-2020 to 10-01-2020',
                          18: '10-01-2020 to 11-01-2020',
                          19: '04-01-2019 to 05-01-2019',
                          20: '05-01-2019 to 06-01-2019',
                          21: '06-01-2019 to 07-01-2019',
                          22: '07-01-2019 to 08-01-2019',
                          23: '08-01-2019 to 09-01-2019',
                          24: '09-01-2019 to 10-01-2019'}})

fig = px.line(df, x="Date", y="Diff", color = 'POD')

import plotly.graph_objects as go
included = 0

lim = {'IAD': {'lower': 90,'upper': 350, 'color':'yellow'},
       'SJCtest': {'lower': 10,'upper': 12, 'color':'green'}}

for i, d in enumerate(fig.data):
    for j, y in enumerate(d.y):
        if y < lim[d.name]['lower'] or y > lim[d.name]['upper']:
             
             if j == 0:
                fig.add_traces(go.Scatter(x=[fig.data[i]['x'][j]],
                                          y=[fig.data[i]['y'][j]],
                                          mode = 'markers',
                                          marker = dict(color=lim[d.name]['color']),
                                          name = d.name + ' threshold',
                                          legendgroup = d.name + ' threshold'))
                included = included + 1
             else:
                fig.add_traces(go.Scatter(x=[fig.data[i]['x'][j-1], fig.data[i]['x'][j]],
                                          y=[fig.data[i]['y'][j-1], fig.data[i]['y'][j]],
                                          mode = 'lines',
                                          # marker = dict(color='yellow'),
                                          line = dict(width = 6, color = lim[d.name]['color']),
                                          name = d.name + ' threshold',
                                          legendgroup = d.name + ' threshold',
                                          showlegend = False if included > 0 else True,
                                         ))
                included = included + 1
            
fig.show()

【讨论】：

嗨@Vestland，现在试试这个。可以说我有 9 个差异。我的完整数据集中的类别。除了修改 lim 之外，我还需要添加什么不同的东西吗？我将更新完整的数据集，它太大了，我不确定该怎么做。我再看看你的笔记。现在试试这个。
@Lynnette 让我们在chat谈一谈
好的@vestland - 实际上我已经创建了一个新帖子。可以回复一下吗？
@Lynnette 刚刚做了。我想。希望这是正确的帖子=）