经济学人风格图表的Python-Seaborn和R-ggplot2绘制

DataCharm

发布于 2021-02-22 15:14:01

8130

发布于 2021-02-22 15:14:01

文章被收录于专栏：数据学术商业新闻数据学术商业新闻

01. 引言

最近着手准备使用R-ggpot2进行图表绘制，当然，Python可视化图表绘制也不能落下，所以，后面的推文我尽量会推出两种教程

。得益于ggplot2丰富的拓展包，本期推文就绘制一幅经济学人风格图表的绘制教程。

02. ggplot2 可视化绘制

在可视化部分，数据处理我们就相对弱化，后面会推出系列的教程的。本次推文所需数据如下（部分）：

绘图代码如下：

library(tidyverse)
library(ggthemes)
library(cowplot)

p <- df_research %>% 
  ggplot(aes(percent_women_ed, field)) + 
    geom_vline(xintercept = 0, color = "black", size = 0.7) + 
    geom_vline(xintercept = 50, color = "red1", size = 0.7) + 
    geom_point(data = filter(df_research, !is.na(type)), aes(color = type, fill = type), 
               size = 9, alpha = 0.8, shape = 21) +
    geom_point(data = filter(df_research, is.na(type)), size = 4.5, 
               color = "black", fill = "grey40", alpha = 0.4, shape = 21) + 
    scale_x_continuous(limits = c(0, 65), breaks = seq(0, 60, by = 10), expand = c(0, 0)) +
    scale_color_manual(values = c("firebrick4", "turquoise4", "dodgerblue3"), name = NULL) +
    scale_fill_manual(values = c("firebrick4", "turquoise4", "dodgerblue3"), name = NULL) +
    guides(color = guide_legend(override.aes = list(size = 4))) +
    labs(x = NULL, y = NULL, 
         caption = '\nSources: "Gender in the Global Research Landscape" by Elsevier') +
    #使用经济学人主题
    theme_economist() +
    theme(text = element_text(family = "Open_Sans"), 
          axis.text = element_text(size = 14),
          axis.ticks.x = element_blank(),
          axis.line.x = element_blank(),
          legend.text = element_text(size = 11),
          legend.position = "top",
          plot.caption = element_text(color = "grey40"),
          plot.background = element_rect(fill = "#dcf0f7"),
          panel.grid.major.y = element_blank(),
          panel.grid.major.x = element_line(color = "grey70", size = 0.4),
          panel.background = element_rect(fill = "#dcf0f7"))
p

当然，我们这里使用cowplot 包的ggdraw()、draw_text()添加一些文本要素。具体如下：

p_research <- ggdraw(p) + 
   draw_text("Still a man's world", x = 0.02, y = 0.98, 
            hjust = 0, vjust = 1, size = 20, family = "Open_Sans_ExtraBold") +
   draw_text("Women among researchers with papers published", 
            x = 0.02, y = 0.93, hjust = 0, vjust = 1, size = 14, family = "Open_Sans") +
   draw_text("(indexed in Scopus from 2011 to 2015, % of total)", 
            x = 0.02, y = 0.88, hjust = 0, vjust = 1, size = 11, family = "Open_Sans")
p_research

这里在绘图系统里进行了数据选择，分别如下：

data = filter(df_research, !is.na(type))
data = filter(df_research, is.na(type))

代码也很明确，即筛选空值和非空值。最终的可视化结果如下：

可以看到，ggplot2 绘制不同风格的主题非常方便，调用包即可，这一点Python可视化绘制则繁琐一点，需一点点绘制精修。

03. Seaborn 可视化绘制

这里使用Python-seaborn 进行绘制可以免去很多繁琐的步骤，作者我也是在尝试使用 matplotlib绘制无果的情况下直接使用seaborn绘制，直接上代码：

#数据筛选
plt.rcParams['font.family'] = ['Open Sans']
#开始绘图
fig,ax = plt.subplots(figsize=(4,4),dpi=200,facecolor='#DCF0F7',edgecolor='#DCF0F7')
ax.set_facecolor("#DCF0F7")
#添加竖线
ax.axvline(x=0,color='k',lw=.9,zorder=0)
ax.axvline(x=50,color='#FF0000',lw=1.5,zorder=0)
ax.invert_yaxis() 
#颜色设置
palette = ['#9B4546','#2C9BA1','#3F8DD5']

scatter = sns.scatterplot(x='percent_women_ed',y='field',hue='country',
                          data=df_research[df_research['type'].isin(['Japan','EU28','Portugal'])],
                          palette=palette,s=400,
                          ec='none',alpha=.95,zorder=1,ax=ax)
scatter_na = sns.scatterplot(x='percent_women_ed',y='field',color='gray',s=100,ec='k',linewidth=.4,alpha=.4
                             data=df_research[~df_research['type'].isin(['Japan','EU28','Portugal'])],
                             ax=ax,zorder=2)
#定制化绘制
ax.tick_params(left=False,bottom=False,labelsize=13,pad=10)
for spine in ['top','bottom','left','right']:
    ax.spines[spine].set_color("none")
ax.grid(which='major',axis='x',ls='-',c='gray',alpha=.6,lw=.5)
ax.set_xlabel("")
ax.set_ylabel("")
ax.set_xticks(np.arange(0,70,10))
ax.set_xlim(left=0,right=65)
ax.set_axisbelow(True)
#修改图例
legend = ax.legend(frameon=False,ncol=4,markerscale=1.5,loc='upper right',fontsize=10,
                   bbox_to_anchor=(1, 1.15),columnspacing=.2)
#去除图例标题
legend.get_texts()[0].set_text('')

ax.text(-.62,1.3, "Still a man's world",transform = ax.transAxes,
        ha='left', va='center',fontsize = 18,color='black',fontweight='bold')
ax.text(-.62,1.2, "Women among researchers with papers published\n(indexed in Scopus from 2011 to 2015, % of total)",
        transform = ax.transAxes,ha='left', va='center',fontsize = 12,color='black')

ax.text(.91,-.12,'\nVisualization by DataCharm',transform = ax.transAxes,
        ha='center', va='center',fontsize = 6,color='black')

plt.savefig(r'F:\DataCharm\Artist_charts_make_python_R\women_research\awomen_research_seaborn.png',
            width=5,height=5,dpi=900,bbox_inches='tight',facecolor='#DCF0F7')

知识点：

（1）数据筛选：

df_research[df_research['type'].isin(['Japan','EU28','Portugal'])]
#和
df_research[~df_research['type'].isin(['Japan','EU28','Portugal'])]

这个步骤筛选出在特定字符串列表的行，也是数据操作中较常使用的方法，大家可以参看下，当然，也可以使用 str.contains()方法进行数据匹配。除此之外，这里使用了 ~ 符号进行“反选”操作。

（2）坐标轴类别数据设置

这里的y轴为具体的分类数据，如下：

matplotlib绘制较为麻烦，而使用seaborn则可完美解决，体现出sns.scatterplot() 的集成优势

。

（3）图例的设置

#修改图例
legend = ax.legend(frameon=False,ncol=4,markerscale=1.5,loc='upper right',fontsize=10,
                   bbox_to_anchor=(1, 1.15),columnspacing=.2)
#去除图例标题
legend.get_texts()[0].set_text('')

seaborn图例设置可是花费我大把时间