我正在做趋势分析，并尝试使用条形图来可视化标签在不同年份的频率。所以我可以看到前 3 个最常见的主题标签术语，并了解这些术语的频率在这些年中是如何演变的。我有一个这样的数据集：
terms year
1 #A;#B;#C 2017
2 #B;#C;#D 2016
3 #C;#D;#E 2021
4 #D;#E;#F 2020
5 #E;#F;#G 2020
6 #F;#G;#H 2020
7 #G;#H;#I 2019
8 #H;#I;#J 2018
9 #I;#J;#K 2020
10 #J;#K;#L 2020
谢谢！
【问题讨论】：
基本上，我们需要计算每年的主题标签。由于特定年份的标签是单列的，我们需要将它分成不同的列，然后我们可以将 df 转换为长 df，我们可以根据年份和标签对其进行分组以查找计数.
library(tidyverse)
structure(list(terms = c("#A;#B;#C", "#B;#C;#D", "#C;#D;#E",
"#D;#E;#F", "#E;#F;#G", "#F;#G;#H", "#G;#H;#I", "#H;#I;#J", "#I;#J;#K",
"#J;#K;#L"), year = c(2017, 2016, 2021, 2020, 2020, 2020, 2019,
2018, 2020, 2020)), row.names = c(NA, -10L), class = c("tbl_df",
"tbl", "data.frame")) -> df
df %>%
separate(terms, into = paste0("t", 1:3), sep = ";") %>%
pivot_longer(-year) %>%
group_by(year, value) %>%
count(value) %>%
ggplot(aes(x = year, y = n, fill = value, label = n)) +
geom_col(position = position_dodge()) +
geom_text(position = position_dodge(1))
由reprex package (v0.3.0) 于 2021-02-05 创建
【讨论】：
NA
要为Base R 中的每一年生成一个可读性好的图表，我们可以执行以下操作：
代码
# First create a list of data.frames that we can utilize to plot
# Split by year
listdf <- split(df, df$year)
# Only choose trends and name list according to year
listdf <- lapply(listdf, "[[", 1)
# Stringsplit to obtain trends as vector for each year
listdfplot <- lapply(listdf, function(x){
unlist(strsplit(x, ";"))
})
# Plot
# Plot side by side
par(mfrow = c(2, 3))
# A barplot for each year
Map(function(x, y){
barplot(table(x), main = paste("Trends in", y), las = 2)
},
listdfplot,
names(listdfplot))
数据
df <- structure(list(terms = c("#A;#B;#C", "#B;#C;#D", "#C;#D;#E",
"#D;#E;#F", "#E;#F;#G", "#F;#G;#H", "#G;#H;#I", "#H;#I;#J", "#I;#J;#K",
"#J;#K;#L"), year = c(2017, 2016, 2021, 2020, 2020, 2020, 2019,
2018, 2020, 2020)), row.names = c(NA, -10L), class = c("tbl_df",
"tbl", "data.frame"))
【讨论】：
如果每个主题标签都有一年的唯一信息，您可以创建一个新的数据框。
之后，您可以使用 geom_bar 处理数据。
因为是新账号，所以无法上传剧情。
library(tidyverse)
library(data.table)
#your Data:
#terms year
#1 #A;#B;#C 2017
#2 #B;#C;#D 2016
#3 #C;#D;#E 2021
#4 #D;#E;#F 2020
#5 #E;#F;#G 2020
#6 #F;#G;#H 2020
#7 #G;#H;#I 2019
#8 #H;#I;#J 2018
#9 #I;#J;#K 2020
#10 #J;#K;#L 2020
# make a df that looks like your data:
terms<- c("#A;#B;#C",
"#B;#C;#D",
"#C;#D;#E",
"#D;#E;#F",
"#E;#F;#G",
"#G;#H;#I",
"#H;#I;#J",
"#I;#J;#K",
"#J;#K;#L")
terms<-as.data.frame(terms)
year<-c(2017,2016,2021,2020,2020,2019,2018,2020,2020)
year<-as.data.frame(year)
df<-cbind(terms,year)
# read your data from what I assume is your Data frame
terms<-c(df$terms)
year.list<-c(df$year)
loopcount<-length(terms)
# make new dummys
year<-c()
hashtags<-c()
all.years<-as.data.frame(hashtags,year)
#split hashtags based on ";"
hashtag.list<-str_split(terms, ";")
通过这个循环，你可以创建一个新的 DF
# make new df were every hashtags gets the information for year
for (i in 1:loopcount){
hashtags<-hashtag.list[[i]]
hashtags<-as.data.frame(hashtags)
year<-c()
for(k in 1:nrow(hashtags)) {
year[k]<- year.list[i]
}
year<-as.data.frame(year)
one.year<-cbind(hashtags,year)
all.years<-rbind(all.years,one.year)
}
hashtagDF<-all.years
head(hashtagDF)
然后可以使用新的 DF 来绘制您想要的内容
或者 如果我理解你是正确的
如果显示每年主题标签的频率，您可以创建一个新的 df 并且只包含前 3 个主题标签
#only include the three most used hashtags per year
# dummys for new df
hashtags<-c()
year<-c()
Freq<-c()
top.3<-as.data.frame(hashtags,year,Freq)
years.in.study<-unique(hashtagDF$year)
#i<-3
for ( i in 1: length(years.in.study)){
what.year<-paste(years.in.study[i])
#subset per year
one.subset<-subset(hashtagDF, year == what.year)
# calculate how often a hashtag is present per year
freq<-table(one.subset)
frequency.per.year<-as.data.frame(freq)
frequency.per.year<-frequency.per.year[order(-frequency.per.year[,3]), ]
# only keep the 3 most occurring terms
lenght.of.file.to.delete<-nrow(frequency.per.year)
if (nrow(frequency.per.year) == 3){
lenght.of.file.to.delete<-lenght.of.file.to.delete+1
}
frequency.per.year<-frequency.per.year[-c(4:lenght.of.file.to.delete), ]
# make a df with all years
top.3<-rbind(top.3,frequency.per.year)
}
top.3
#order for year
top.3$year<-as.character(top.3$year)
top.3[order(top.3[,2]), ]
#year should be a factor
top.3$yearF<-as.factor(top.3$year)
然后你可以绘制它
# plot as barplot
# with
# the frequencies of the hashtags in different years.
# the top 3 most frequent hashtag terms per year
barplot<-ggplot(data=top.3, aes(x=yearF, y=Freq,fill=hashtags)) +
geom_bar(stat="identity")+
labs(title = "",
subtitle = "",
caption = "",
x= "",
y= "")
barplot
ggsave(file="hashtag.png", barplot, width = 210, height = 297, units = "mm")
【讨论】：