第 2 章 学科结构与发展

2.1 PROS

关于文献计量的数据库很多，比如PROS，Arxiv等等，对PROS和Arxiv数据集的分析，另外单独介绍。

2.2 Web of Science

我们选择Web of Science，原因是因为当前ESI学科是各高校关注的热点，而Web of Science数据库是ESI的上游数据库，挖掘的空间更大。

图 2.1: ESI与Web of Science的关系图.

Web of Science数据库的题录数据结构如下表：

2.3 数据集的构建

由于wos网站限制每次只能下载500条记录，因此每个学校的题录文件（一个文件夹）存放不止一个txt文件。具体文件结构如下：

除了题录数据，还包括辅助的数据集

编号	名称	简介	变量数量	共有变量
1	web of science	wos题录	13	University,ISSN
2	sichuan_shool_encn2.csv	四川高校的名称	2	University
3	esi_plus_cas_IF_set.rds	esi期刊学科分类	11	ISSN
		中科院学科分类
		是否中国SCI期刊
		期刊影响因子等

接下来，我们来读取规整文件

# Load Packages
library(tidyverse)
library(here)
library(fs)
library(knitr)
library(kableExtra)

构建读取相应文件的子函数

read_plus <- function(flnm) {
        read_tsv(flnm, quote = "", col_names = TRUE) %>% 
        #or
        #read_delim(flnm,  delim="\t" , quote = "", col_names = TRUE) %>% 
        select(AU, AF, SO, DE, C1, RP, FU, CR, TC, SN, PY, UT) %>% 
        mutate(University = flnm %>%                 # 加入了学校名
                            str_split("/", simplify = TRUE) %>%
                            .[, length(.) - 1] %>%   # 倒数第二列
                            str_sub(start = 4)
               ) 
}

"G:/Guide_DS_R/63R4IS/data/Record/sichuan/01_Chengdu_Univ_Technol/1-500.txt" %>% 
   str_split("/", simplify = F) %>% 
   map_chr(.,
     ~.x[length(.x) - 1]
     )

## [1] "01_Chengdu_Univ_Technol"

依次迭代读取

tbl <- here("data", "Record", "sichuan") %>% 
  dir_ls(regexp = "*.txt", recursive = TRUE) %>%  
  map_dfr(~read_plus(.))
tbl

测试下呢

tbl %>% count(University)

读取四川高校的名称

sichuan_univ_name <- read_csv(here("data", "map_list",  "sichuan_shool_encn2.csv") )
sichuan_univ_name

合并到tbl， 形成 sichuan_set

sichuan_set <- tbl %>% 
               left_join(sichuan_univ_name, by = "University")

这里是esi期刊列表数据集， esi_plus_cas_set = esiJCR + CAS + chineseJCR （即，每个期刊的esi学科分类，中科院学科分类，是否是中国JCR期刊），目的是方便今后分析

esi_plus_cas_IF_set <- read_rds(here("data", "esiJournalsList",  "esi_plus_cas_IF_set.rds"))

将 esi_plus_cas_IF_set 合并到 sichuan_set， 形成 complete_set

complete_set <- left_join(sichuan_set, esi_plus_cas_IF_set, by = c("SN" = "ISSN") ) %>% 
               select(University, University_cn, everything()) %>%
               rename(ISSN = SN) %>% 
               mutate(SO = str_to_title(SO))

2.4 数据集的导入

最终，我们围绕四川省各高校，选择Web of Science数据库中的10个维度（上图中红色圈出的部分），然后映射并融合了ESI学科分类、中科院期刊分类、期刊等级以及影响因子等，构建了包含24个维度的完整数据集。

#complete_set <- read_rds("data/dataset/complete_set.rds")
glimpse(complete_set)

## Rows: 19,341
## Columns: 24
## $ University      <chr> "Chengdu_Univ_Technol", "Chen~
## $ University_cn   <chr> "成都理工大学", "成都理工大学~
## $ AU              <chr> "Duan, YP; Shang, RS; Zhong, ~
## $ AF              <chr> "Duan, Yiping; Shang, Ruishu;~
## $ SO              <chr> "International Journal Of Hyd~
## $ DE              <chr> "Dry reforming of methane; Hy~
## $ C1              <chr> "[Duan, Yiping; Shang, Ruishu~
## $ RP              <chr> "Huang, LH (reprint author), ~
## $ FU              <chr> "National Natural Science Fou~
## $ CR              <lgl> NA, NA, NA, NA, NA, NA, NA, N~
## $ TC              <dbl> 0, 0, 0, 0, 0, 0, 0, 1, 0, 0,~
## $ ISSN            <chr> "0360-3199", "1367-9120", "09~
## $ PY              <dbl> 2016, 2016, 2016, 2016, 2016,~
## $ UT              <chr> "WOS:000389786500012", "WOS:0~
## $ Full_title      <chr> "INTERNATIONAL JOURNAL OF HYD~
## $ Category_ESI    <chr> "ENGINEERING", "GEOSCIENCES",~
## $ Category_ESI_cn <chr> "工程学", "地球科学", "计算机~
## $ Category_CAS    <chr> "工程技术", "地学", "工程技术~
## $ Degree_CAS      <int> 2, 3, 2, 4, 4, 4, 3, 3, 3, 4,~
## $ Title           <chr> NA, NA, NA, NA, NA, NA, "石油~
## $ is_chinese      <dbl> 0, 0, 0, 0, 0, 0, 1, 1, 1, 1,~
## $ Cites           <dbl> 75123, 12551, 22735, 696, 890~
## $ ImpactFactor    <dbl> 4.229, 2.866, 3.241, 0.314, 2~
## $ Eigenfactor     <dbl> 0.08137, 0.02432, 0.04892, 0.~

2.5 各高校对比

建立数据集有很大的好处，比如我们可以很方便的对比四川各高校的科研情况

complete_set %>% count(University_cn) %>% arrange(desc(n))

被引情况呢

complete_set %>% count(University_cn, wt = TC) %>% arrange(desc(n))

从未被引用的数量和占比（太讨厌了吧）

complete_set %>% group_by(University_cn) %>%
                 summarise(num = n(),
                            not_cited = sum(TC == 0),
                            prop = not_cited/num)  %>% 
                 arrange(desc(prop))

2.6 分析单个学校

看看我们川师的情况吧

sicnu_set <- complete_set %>% filter(University == "Sichuan_Normal_Univ")  
sicnu_set %>% summarise(count = n()) 

sicnu_set %>%  
      filter(str_detect(C1, "Sichuan Normal Univ,\\s+([^,]*),")) %>% 
      summarise(count = n()) 

为什么少了2各呢？大家想想吧

2.7 第一作者筛查

sicnu_set %>%
      filter(str_detect(C1, "^\\[([^]]+)\\]\\s+Sichuan Normal Univ,\\s+([^,]*),")) %>% 
      summarise(count = n()) 

2.8 通讯作者筛查

sicnu_set %>%
     filter(str_detect(RP, "Sichuan Normal Univ")) %>% 
     summarise(count = n()) 

2.9 可视化之小试牛刀

川师逐年论文发表情况

sicnu_set %>% ggplot() +
              geom_bar(aes(x = PY))

2.10 可视化之初显身手

some_univ <- c("四川师范大学", "成都理工大学", "西南科技大学", 
               "成信大", "西南石油大学", "四川农业大学", 
               "四川理工学院", "西华师范大学","西华大学")
some_cate <- c("物理学", "化学", "数学", "工程学", "计算机科学", "材料科学")

library(showtext)
showtext_auto()
complete_set  %>%
  group_by(University_cn, Category_ESI_cn, PY) %>%
  summarise(sumpaper = n()) %>%
  filter(University_cn %in% some_univ,
          Category_ESI_cn %in% some_cate, 
          PY >= 2010
          ) %>%
   ggplot(aes(x = PY, y = sumpaper, group = University_cn, colour = factor(University_cn)) ) + 
   geom_point() +
   geom_line() +
   facet_wrap(vars(Category_ESI_cn), scales = "free") +
   labs(x = NULL, y = NULL, colour = "") +
   ggtitle("我省部分省属高校近几年各学科论文情况", subtitle = '（数据来源 Web of Science）') + 
  guides(fill = guide_legend( nrow = 1, byrow = TRUE)) +
  theme(
     #plot.title = element_text(size = 16),
     #plot.subtitle=element_text(size=14, hjust=0.01, face="italic"),
     legend.box = "horizontal",
     legend.position = "bottom",
     legend.direction = "horizontal"#, 
     #text=element_text(size=14)
     )

complete_set %>%
  filter(!is.na(Category_ESI_cn)) %>% 
  group_by(University_cn, Category_ESI_cn) %>%
  summarise(sumpaper = n()) %>%
  filter(University_cn %in% some_univ) %>%
  ggplot(aes(x = University_cn, y = Category_ESI_cn, colour = University_cn, fill = University_cn, size = sumpaper)) + 
  geom_point(shape = 21 ) +
  scale_size(range = c(0,20)) +
  xlab("") + ylab("") + 
  ggtitle("我省部分省属高校近几年各学科论文情况", subtitle = '（数据来源 Web of Science）') + 
  theme(legend.position = "none")

2.11 可视化之渐入佳境

library(scales)
library(ggiraph)
library(ggiraphExtra)

df_radar <- complete_set %>% 
     filter(!is.na(Category_ESI_cn)) %>% 
     count(University_cn, Category_ESI_cn) %>% 
     spread(Category_ESI_cn, n, fill = 0) %>% 
     mutate_at(vars(-University_cn), funs(rescale)) %>%
     filter(University_cn %in% some_univ) 

ggRadar(data = df_radar, 
        mapping = aes(colour = University_cn),
        size = 2,
        legend.position = "none",
        rescale = TRUE) +
  facet_wrap(vars(University_cn) )

来点动画吧，(gganimate升级后，该代码无法工作)

现在我们可以使用新的 gganimate宏包了

library(gganimate)
ttb <- complete_set %>% 
       filter(University_cn %in% some_univ, !is.na(Category_ESI_cn)) %>% 
       count(University_cn, PY, Category_ESI_cn) #%>%

看看这个动画

ttb %>% 
  ggplot(aes(x = University_cn, y = Category_ESI_cn, colour = University_cn, fill = University_cn, size = n)) + 
  geom_point(shape = 21 ) +
  scale_size(range = c(0,10)) +
  # Here comes the gganimate code
  labs(title = "{closest_state}") +
  transition_states(PY, state_length = 2, transition_length = 1) +
  enter_fade() + 
  exit_shrink() +
  ease_aes('linear') + 
  xlab("") + ylab("") + 
  theme(legend.position = "none")

又来一个

tsb <- complete_set %>% 
     count(University_cn, PY) %>%
     filter(University_cn %in% some_univ)
 

 
tsb %>% 
  ggplot(aes(PY, n, size = n, colour = University_cn)) +
  #geom_point(alpha = 0.7, show.legend = FALSE) +
  geom_point(alpha = 0.7, show.legend = TRUE) +
  scale_size(range = c(2, 12)) +
  # Here comes the gganimate specific bits
  labs(title = "我省部分省属高校近几年各学科论文情况Year: {frame_time}",
          subtitle = '（数据来源 Web of Science）',
          y = "", x = "Years",
          caption = "Source: Web of Science",
          colour = "学校") +
  transition_time(PY) +
  shadow_mark(past = TRUE) +  #equivalent to cumulative= TRUE in the old package
  ease_aes('linear')

再来一个

library(RColorBrewer)
library("viridis")

mycols <- brewer.pal(9,"Oranges")[2:8]  

theme_set(theme_bw())

shapenumber %>% 
   filter(!is.na(Category_CAS)) %>%
   ggplot(aes(y = Category_CAS, x = Degree_CAS, fill = value)) + 
   geom_tile(color = "white") +
   scale_fill_gradientn(colours = mycols) +
   geom_text( aes(label = round(value,2)),
              size = 6
             ) +
   labs(x = NULL, y = NULL) +
   ggtitle("四川师范大学近几年各学科论文发表情况", subtitle = '（数据来源 Web of Science）') + 
   guides(fill = guide_legend(
                     title = "数量"
                     #reverse = TRUE
                    )
          ) +
   theme(
     plot.title = element_text(size = 18),
     plot.subtitle = element_text(size = 14, hjust = 0.01, face = "italic"),
     legend.title = element_text(size = 14),
     legend.position = "top",
     legend.justification = c(1,0),
     legend.margin = margin(0, 0, 0, 0),
     axis.text.x = element_text(margin = margin(0.4, unit = "cm")),
     text = element_text(family = "fzxbsj",size = 20)
     #plot.margin = unit(c(0, 0, 0, 0), "cm") #top, right, bottom, left
     ) +  
     scale_x_discrete(expand = c(0,0), 
                      labels = c("degree1" = "一区", "degree2" = "二区", 
                                 "degree3" = "三区", "degree4" = "四区")
                     ) +
    scale_y_discrete(expand = c(0,0)) 

2.12 更多

还可以做各种统计和可视化，在后面的章节一一展示吧