是否可以在 ggplot2 的单个时间序列图中绘制多条趋势线？答案

【问题标题】：Is it possible to draw multiple trendlines within a single time series graph in ggplot2?是否可以在 ggplot2 的单个时间序列图中绘制多条趋势线？
【发布时间】：2021-02-18 02:56:01
【问题描述】：

我计划在一个时间序列图中绘制几条趋势线，以便我可以可视化跨越多个时间限制的趋势变化。我设法在整个时间序列中绘制了一个线性趋势，但我希望可能在 2007 年至 2010 年和 2010 年至 2013 年绘制更多趋势，这将有助于我解决 2007 年至 2010 年之间的稳定趋势和 2010 年之间的下降模式和 2013 年。我使用了以下代码：

data <- read.csv("sample.csv",
                 header = T,
                 sep = ",",
                 dec = ".")
head(data)
data$Year <- as.Date(data$ï..Date, format = "%m/%d/%Y")
class(data$Year)
attach(data)
time_plot <- ggplot(data, aes(x = Year, y = SPM)) +
  geom_line(color = 'black', size = 1.3)  + geom_point(color = "blue", size = 1.3) +
  scale_x_date(date_labels = "%Y", date_breaks = "1 year") + xlab(label = "Time (Years)") + ylab(label = "Concentration") +
    theme_bw() + stat_smooth(
      method = "lm",
      formula = y ~ x,
      size = 0.75,
      se = T,
      color = "blue",
      fill = "#9AE5D7"
    ) + stat_poly_eq(
      face = "bold",
      parse = T,
      aes(label = ..eq.label..),
      formula = y ~ x,
      label.x.npc = 0.5,
      label.y.npc = 0.1,
      size = 6,
      coef.digits = 4
    ) +
    theme(
      plot.title = element_text(
        size = 17,
        face = "bold",
        colour = "black"
      ),
      axis.title.x = element_text(
        size = 20,
        face = "bold",
        colour = "black"
      ),
      axis.title.y = element_text(
        size = 20,
        face = "bold",
        colour = "black"
      ),
      axis.text.x = element_text(
        size = 18,
        face = "bold",
        colour = "black"
      ),
 
      axis.text.y = element_text(
        size = 18,
        face = "bold",
        colour = "black"
      ),
     
      strip.text.x = element_text(
        size = 16,
        
        face = "bold",
        colour = "black"
      ),
      strip.text.y = element_text(
        size = 16,
        
        face = "bold",
        colour = "black"
      ),
      axis.line.x = element_line(color = "black", size = 1),
      axis.line.y = element_line(color = "black", size = 1),
      axis.ticks = element_line(color = "black", size = 1.2),
      axis.ticks.length = unit(0.2, "cm"),
      panel.border = element_rect(
        colour = "black",
        fill = NA,
        size = 1
      ),
      legend.title = element_blank(),
      legend.position = c(.8, .2),
    ) +
    stat_fit_glance(
      method = 'lm',
      method.args = list(formula = y ~ x),
      geom = 'text',
      aes(label = paste(
        "P-value = ", signif(..p.value.., digits = 4), sep = ""
      )),
      size = 6,
      label.x = "left",
      label.y = "top",
    )

返回如下图：

但是，我希望生成这样的图，其中包含几条趋势线：

之前在堆栈溢出中有一个帖子，它是similar to my query，但它是针对“python”的。我在想我是否可以在 R 中使用 ggplot2 做类似的事情？如果您能花点时间为我的问题提供一些解决方案，或者建议任何可以帮助我生成此类数据的教程、网站或软件包，我将不胜感激。我还可以访问黄金软件的绘图员，这是获得此类数据的更好平台吗？我附上下面的数据集：

ï..Date   SPM       Year
1   1/1/2007 6.412 2007-01-01
2   2/1/2007 7.827 2007-02-01
3   3/1/2007 6.816 2007-03-01
4   4/1/2007 8.021 2007-04-01
5   5/1/2007 6.130 2007-05-01
6   6/1/2007 4.982 2007-06-01
7   7/1/2007 3.776 2007-07-01
8   8/1/2007 4.764 2007-08-01
9   9/1/2007 5.699 2007-09-01
10 10/1/2007 7.264 2007-10-01
11 11/1/2007 8.168 2007-11-01
12 12/1/2007 7.518 2007-12-01
13  1/1/2008 7.157 2008-01-01
14  2/1/2008 7.996 2008-02-01
15  3/1/2008 5.806 2008-03-01
16  4/1/2008 5.388 2008-04-01
17  5/1/2008 6.535 2008-05-01
18  6/1/2008 3.715 2008-06-01
19  7/1/2008 4.723 2008-07-01
20  8/1/2008 4.259 2008-08-01
21  9/1/2008 5.524 2008-09-01
22 10/1/2008 7.755 2008-10-01
23 11/1/2008 8.393 2008-11-01
24 12/1/2008 5.702 2008-12-01
25  1/1/2009 5.816 2009-01-01
26  2/1/2009 5.954 2009-02-01
27  3/1/2009 5.142 2009-03-01
28  4/1/2009 6.286 2009-04-01
29  5/1/2009 7.408 2009-05-01
30  6/1/2009 5.866 2009-06-01
31  7/1/2009 7.188 2009-07-01
32  8/1/2009 3.729 2009-08-01
33  9/1/2009 4.284 2009-09-01
34 10/1/2009 6.392 2009-10-01
35 11/1/2009 6.642 2009-11-01
36 12/1/2009 6.365 2009-12-01
37  1/1/2010 6.999 2010-01-01
38  2/1/2010 6.906 2010-02-01
39  3/1/2010 6.205 2010-03-01
40  4/1/2010 3.497 2010-04-01
41  5/1/2010 2.278 2010-05-01
42  6/1/2010 3.510 2010-06-01
43  7/1/2010 4.112 2010-07-01
44  8/1/2010 5.469 2010-08-01
45  9/1/2010 5.402 2010-09-01
46 10/1/2010 5.365 2010-10-01
47 11/1/2010 6.412 2010-11-01
48 12/1/2010 7.384 2010-12-01
49  1/1/2011 7.660 2011-01-01
50  2/1/2011 7.380 2011-02-01
51  3/1/2011 7.880 2011-03-01
52  4/1/2011 5.971 2011-04-01
53  5/1/2011 6.944 2011-05-01
54  6/1/2011 3.911 2011-06-01
55  7/1/2011 4.438 2011-07-01
56  8/1/2011 3.266 2011-08-01
57  9/1/2011 4.554 2011-09-01
58 10/1/2011 7.247 2011-10-01
59 11/1/2011 7.607 2011-11-01
60 12/1/2011 7.791 2011-12-01
61  1/1/2012 9.193 2012-01-01
62  2/1/2012 7.312 2012-02-01
63  3/1/2012 3.753 2012-03-01
64  4/1/2012 3.458 2012-04-01
65  5/1/2012 1.211 2012-05-01
66  6/1/2012 2.052 2012-06-01
67  7/1/2012 2.055 2012-07-01
68  8/1/2012 3.804 2012-08-01
69  9/1/2012 5.728 2012-09-01
70 10/1/2012 6.501 2012-10-01
71 11/1/2012 5.177 2012-11-01
72 12/1/2012 4.829 2012-12-01

任何帮助、建议或建议将不胜感激。提前致谢。

【问题讨论】：

标签： r ggplot2 time-series data-visualization trendline

【解决方案1】：

您可以简单地使用原始数据框的子集重复您的 geom_smooth 调用：

ggplot(data, aes(x = Year, y = SPM)) +
  geom_line(color = 'black', size = 1.3)  + 
  geom_point(color = "blue", size = 1.3) +
  stat_smooth(method = "lm", formula = y ~ x, size = 0.75, se = TRUE,
              color = "blue", fill = "#9AE5D7") + 
  stat_smooth(method = "lm", formula = y ~ x, size = 0.75, se = TRUE,
              color = "red", fill = "red", alpha = 0.2, 
              data = data[data$Year < as.Date("2009-06-01"),]) + 
  stat_smooth(method = "lm", formula = y ~ x, size = 0.75, se = TRUE,
              color = "forestgreen", fill = "forestgreen", alpha = 0.2,
              data = data[data$Year >= as.Date("2009-06-01"),]) + 
  stat_poly_eq(face = "bold", parse = TRUE, aes(label = ..eq.label..),
               formula = y ~ x, label.x.npc = 0.5, label.y.npc = 0.1, 
               size = 6, coef.digits = 4) +
  stat_fit_glance(method = 'lm', method.args = list(formula = y ~ x),
                  geom = 'text', 
                  aes(label = paste(
                    "P-value = ", signif(..p.value.., digits = 4), sep = ""
                  )), size = 6, label.x = "left", label.y = "top") +
  scale_x_date(date_labels = "%Y", date_breaks = "1 year") + 
  labs(x = "Time (Years)", y = "Concentration") +
  theme_bw() + 
  theme(plot.title        = element_text(size = 17, face = "bold"),
        axis.title.x      = element_text(size = 20, face = "bold"),
        axis.title.y      = element_text(size = 20, face = "bold"),
        axis.text.x       = element_text(size = 18, face = "bold"),
        axis.text.y       = element_text(size = 18, face = "bold"),
        strip.text.x      = element_text(size = 16, face = "bold"),
        strip.text.y      = element_text(size = 16, face = "bold"),
        axis.line.x       = element_line(color = "black", size = 1),
        axis.line.y       = element_line(color = "black", size = 1),
        axis.ticks        = element_line(color = "black", size = 1.2),
        axis.ticks.length = unit(0.2, "cm"),
        panel.border      = element_rect(fill = NA, size = 1),
        legend.title      = element_blank(),
        legend.position   = c(.8, .2))

在这种情况下，整体趋势相当稳定，因此背景蓝线被两个部分线段遮挡。

【讨论】：

非常感谢您的调查。但是，您介意我请您解释这部分吗data = data[data$Year < as.Date("2009-06-01"),])。您使用“2009”作为中间值在其两侧绘制趋势线。如果我有一个特定的趋势上限和下限怎么办？例如，我想绘制从 2007 年到 2009 年和 2011 年到 2013 年的趋势。对于这样的事情，我需要在代码中做哪些修改？
@asheef_ik 你可以做到data = data[data$Year >= as.Date("2007-01-01") & data$Year <= as.Date("2009-01-01"),] 和data = data[data$Year >= as.Date("2011-01-01") & data$Year <= as.Date("2013-01-01"),]
问题解决了，再次感谢。现在我已经绘制了趋势，如果我可以将趋势的线类型（如我帖子中的一张图片）更改为虚线或点划线，似乎会更好。我知道这可能与最初的问题无关，你能帮我解决吗？
@asheef_ik 只需将linetype = 2 添加到geom_smooth 调用中。您可以尝试不同的数字，看看哪个合适。