dashboard.Rmd

---
title: "Hybrid OA Journal Monitor"
output: 
  flexdashboard::flex_dashboard:
    navbar:
      - { title: "About", href: "https://subugoe.github.io/hybrid_oa_dashboard/about.html"}
      - { title: "Data", href: "https://github.com/subugoe/hybrid_oa_dashboard/tree/master/data/README.md"}
    orientation: rows
    source_code: "https://github.com/subugoe/hybrid_oa_dashboard/"
    vertical_layout: scroll 
runtime: shiny
---

```{r setup, include=FALSE}
library(flexdashboard)
library(viridis)
library(tidyverse)
library(plotly)
library(shiny)
library(scales)
library(DT)
library(writexl)
library(echarts4r)
```

```{r global}
# access to data
hybrid_df <- readr::read_csv("data/hybrid_publications.csv") %>%
  mutate(license = fct_infreq(license)) %>%
  mutate(publisher = fct_infreq(publisher)) %>%
  mutate(year = factor(issued, levels = c("2013", "2014", "2015","2016", "2017", "2018", "2019"))) %>% 
  mutate(hybrid_type = ifelse(hybrid_type == "SCOAP", "SCOAP<sup>3</sup>", hybrid_type)) %>%
  mutate(hybrid_type = factor(hybrid_type, levels = c("Open APC (TA)", "Open APC (Hybrid)", "SCOAP<sup>3</sup>"))) %>%
  mutate(domain = paste(domain, suffix, sep =".")) %>%
  arrange(desc(yearly_publisher_volume))
unpaywall_df <- readr::read_csv("data/unpaywall_df.csv") %>%
  mutate(year = factor(year, levels = c("2013", "2014", "2015","2016", "2017", "2018", "2019")))
```

Overview
=================================================================

Inputs {.sidebar}
-----------------------------------------------------------------------


```{r}
# from https://stackoverflow.com/questions/42148177/how-can-i-build-multiple-inputs-into-my-shiny-app-with-both-updating-and-multipl

selectizeInput("publisherinput", "Publisher", c("All", levels(hybrid_df$publisher)), selected =
              "All")
selectizeInput("journalinput", "Journal", c("All", unique(hybrid_df$journal_title)), selected =
              "All", options = list(maxOptions = 5000))

# publisher filter
jn_filtered <-
  reactive(hybrid_df[input$publisherinput == "All" |
                       hybrid_df$publisher == input$publisherinput,])

observe(updateSelectInput(
  session,
  "journalinput",
  choices = c("All", sort(unique(jn_filtered()$journal_title))),
  selected = "All"
))

# journal_title filter
jn_f <-
  reactive(jn_filtered()[input$journalinput == "All" |
  jn_filtered()$journal_title == input$journalinput, ])
# prepare dataset for plotting
hybrid_sub <- reactive({
if (length(unique(jn_f()$publisher)) > 1) {
    hybrid_sub <- jn_f() %>%
      group_by(year, yearly_all, license) %>%
      count() %>%
      mutate(prop = n / yearly_all)
  } else if (length(unique(jn_f()$journal_title)) > 1 &&
             length(unique(jn_f()$publisher)) == 1) {
    hybrid_sub <- jn_f() %>%
      group_by(year, license, yearly_publisher_volume) %>%
      count() %>%  
      mutate(prop = n / yearly_publisher_volume)
  } else {
    hybrid_sub <- jn_f() %>%
      group_by(year, license, yearly_jn_volume) %>%
      count() %>%  
      mutate(prop = n / yearly_jn_volume)
  }
})
# using unpaywall indicators
# unpaywall <- reactive({
#   hybrid_df %>%
#     filter(journal_title %in% jn_f()$journal_title) %>%
#     group_by(year, journal_title, publisher, jn_y_unpaywall_others) %>%
#     summarise(n = n_distinct(doi_oa)) %>%
#     gather(n, jn_y_unpaywall_others, key = "source", value = "articles") %>%
#     ungroup() %>%
#     group_by(year, source) %>%
#     summarise(articles = sum(articles, na.rm = TRUE)) %>%
#     mutate(
#       source = ifelse(
#         source == "n",
#         "Crossref immediate license",
#         "Other license information\n(Unpaywall)"
#       )
#     )
# })
```

Notice that only those hybrid open access journals were included where 
academic institutions sponsored the open access publication according to the [Open APC initiative](https://github.com/openapc/openapc-de) 
and where publishers shared license information about immediate open access with [Crossref](https://github.com/CrossRef/rest-api-doc).

Data sources:

<a href="https://github.com/CrossRef/rest-api-doc"><img src="https://assets.crossref.org/logo/metadata-from-crossref-logo-200.svg" width="200" height="68" alt="Metadata from Crossref logo"></a>

<a href= "https://github.com/openapc/openapc-de"><img src="https://www.intact-project.org/public/img/openAPC_01+.png" width="200"  alt="Metadata from open apc logo"></a>

Build by: <a href= "https://www.sub.uni-goettingen.de/sub-aktuell/"><img src="img/sub_logo.svg" width="200"  alt="SUB Göttingen"></a>.

Row
-----------------------------------------------------------------------

### Publishers selected

```{r}
renderValueBox({ 
  publisher_n <- length(unique(jn_f()$publisher))
  valueBox(publisher_n, icon = "fa-filter")
})
```

### Journals selected

```{r}
renderValueBox({
  journal_n <- length(unique(jn_f()$journal_title))
  valueBox(format(journal_n, big.mark = " ", scientific = FALSE), 
           icon = "fa-filter")
})
```

### Hybrid OA articles indexed in Crossref

```{r}
renderValueBox({
  hybrid_n <- nrow(jn_f())
  valueBox(format(hybrid_n, big.mark = " ", scientific = FALSE), 
           icon = "fa-creative-commons")
})
```


Column {.tabset .tabset-fade}
-----------------------------------------------------------------------

### Hybrid OA licenses found (relative)

```{r}
renderPlotly({
  p <- ggplot(hybrid_sub(), aes(year, prop, fill = license)) +
    xlab("Year") +
    ylab("Hybrid OA / Articles published") +
    # palette https://twitter.com/debivort/status/994583058031546369
    scale_fill_manual(
      "License",
      values =  c(
        "#65BADA",
        "#068E8C",
        "#00A757",
        "#E5BA52",
        "#D86F27",
        "#C82E6B"
      )
    ) +
    scale_x_discrete(drop = FALSE) +
    scale_y_continuous(labels = scales::percent) +
    theme_minimal() +
    geom_col(position = position_stack(reverse = TRUE))
  plotly::ggplotly(p)
})
```


### Hybrid OA licenses found (absolute)

```{r}
renderPlotly({
  p <- ggplot(hybrid_sub(), aes(year, n, fill = license)) +
    xlab("Year") +
    ylab("Hybrid OA Articles") +
    # pallette https://twitter.com/debivort/status/994583058031546369
    scale_fill_manual(
      "License",
      values =  c(
        "#65BADA",
        "#068E8C",
        "#00A757",
        "#E5BA52",
        "#D86F27",
        "#C82E6B"
      )
    ) +
    scale_x_discrete(drop = FALSE) +
    scale_y_continuous(
      labels = function(x)
        format(x, big.mark = " ", scientific = FALSE),
      breaks = pretty_breaks()
    ) +
    theme_minimal() +
    geom_col(position = position_stack(reverse = TRUE))
  plotly::ggplotly(p)
})
```

### Other types of OA license information detected by Unpaywall

```{r}
renderPlotly({
  p <- unpaywall_df %>% 
    filter(journal_title %in% jn_f()$journal_title) %>%
    group_by(year, source) %>%
    summarise(articles = sum(articles, na.rm = TRUE)) %>%
  ggplot(aes(year, articles, fill = source)) +
    geom_bar(stat = "identity", position = position_stack(reverse = TRUE)) +
    xlab("Year") +
    ylab("Articles") +
    scale_x_discrete(drop = FALSE) +  
    scale_y_continuous(
      labels = function(x)
        format(x, big.mark = " ", scientific = FALSE),
      breaks = scales::pretty_breaks()
    ) +
    scale_fill_manual(
      "OA articles with",
      values = c(
        "Crossref immediate license" = "#56B4E9",
        "Other license information\n(Unpaywall)" = "#b3b3b3a0"
      )
    ) +
    theme_minimal()
  plotly::ggplotly(p)
})
```

Row {data-width=400 data-height=350}
-------------------------------------

### Sources of disclosure of OA sponsorship

```{r}
renderPlotly({
  p <- hybrid_df %>%
    filter(journal_title %in% jn_f()$journal_title) %>%
    ggplot(aes(year, ..count.., fill = hybrid_type)) +
    geom_bar(position = position_stack(reverse = TRUE)) +
    xlab("Year") +
    ylab("Articles") +
    scale_x_discrete(drop = FALSE) +  
    scale_y_continuous(
      labels = function(x)
        format(x, big.mark = " ", scientific = FALSE),
      breaks = scales::pretty_breaks()
    ) +
    scale_fill_manual(
      "Sources",
      values = c(
        "Open APC (TA)" = "#23445D",
        "Open APC (Hybrid)" = "#AE4132",
        "SCOAP<sup>3</sup>" = "#12AAB5"
      ),
      na.value = "#b3b3b3a0"
    ) +
    theme_minimal()
  
  tt <- plotly::ggplotly(p, tooltip = c("y"))
  tt$x$data <- lapply(tt$x$data, function(x) {
    x$text <- paste(x$name, x$y, sep = ": ")
    x
  })
  tt
})
```


### Text-mined author emails (first match per article)

```{r}
renderEcharts4r({
  hybrid_df %>%
    filter(journal_title %in% jn_f()$journal_title) %>%
    count(suffix, domain) %>%
    filter(!is.na(domain)) %>%
    e_charts() %>%
    e_treemap(suffix, domain, n, leafDepth = "1",
              name = "Corresponding Email") %>%
    e_tooltip(trigger = "item")
})
```

Compare 
=====================================

Row
--------------------------------------------

### Hybrid open access uptake by Publisher, shown as the proportion of the total number of hybrid open articles found

```{r}
hybrid_all <-
  hybrid_df %>%
  count(year) %>%
  mutate(year = gsub("20", "", year))
hybrid_df %>%
  mutate(year = gsub("20", "", year)) %>%
  mutate(publisher_group = fct_lump(publisher, prop = 0.05)) %>%
  group_by(year, publisher_group) %>% 
  summarize(n = n()) %>% 
  mutate(prop = n / sum(n)) %>%
  mutate(`Proportion in %` = round(prop * 100, 2)) %>%
  ggplot(aes(year, n)) +
  geom_bar(
    data = hybrid_all,
    aes(fill = "All Hybrid OA Articles"),
    color = "transparent",
    stat = "identity"
  ) +
  geom_bar(aes(fill = "by Publisher", label = `Proportion in %`), color = "transparent", stat = "identity") +
  facet_wrap(~publisher_group, nrow = 1) +
  scale_fill_manual(values = c("#b3b3b3a0", "#56B4E9"), name = "") +
   labs(x = "Year", y = "Hybrid OA Articles") +
  theme_minimal() +
  theme(legend.position = "top",
        legend.justification = "right") +
  scale_y_continuous(labels = scales::number_format(big.mark = " ")) +
  theme(panel.grid.minor = element_blank()) +
  theme(axis.ticks = element_blank()) +
  theme(panel.grid.major.x = element_blank()) +
  theme(panel.border = element_blank()) -> publishers_comp_plot

plotly::ggplotly(publishers_comp_plot, tooltip = c("label", "y"))
```

Row
--------------------------------------------


### Journal hybrid open access share by publisher 

```{r}
fillCol(
  height = 600,
  flex = c(NA, 1),
  inputPanel(
    selectInput(
      "Year",
      label = "Select Year",
      choices = c(rev(levels(hybrid_df$year))),
      selected = c("2017", "2018"),
      multiple = TRUE
    ),
    checkboxInput("rb", "Display journals", FALSE)
  ),
  plotlyOutput("pubplot", height = "100%")
)

output$pubplot <- renderPlotly({
  if (input$Year == "All") {
    tt <-   hybrid_df %>%
      filter(!is.na(license)) %>%
      mutate(publisher = forcats::as_factor(publisher)) %>%
      group_by(year, publisher, journal_title, yearly_jn_volume) %>%
      summarize(oa = n()) %>%
      mutate(prop = oa / yearly_jn_volume)
  } else {
    tt <-  hybrid_df %>%
      filter(!is.na(license)) %>%
      mutate(publisher = forcats::as_factor(publisher)) %>%
      filter(year %in% input$Year) %>%
      group_by(year, publisher, journal_title, yearly_jn_volume) %>%
      summarize(oa = n()) %>%
      mutate(prop = oa / yearly_jn_volume)
  }
  tt <- tt %>%
    ungroup() %>%
    mutate(publisher = forcats::fct_other(publisher, keep = levels(publisher)[1:10]))
  p <- ggplot(tt, aes(y = prop, x = as.factor(publisher))) +
    geom_boxplot() +
    coord_flip() +
    theme_minimal() +
    xlab("Publisher") +
    ylab(NULL) +
    scale_x_discrete(
      drop = FALSE,
      limits = rev(levels(tt$publisher)),
      labels = function(x)
        str_wrap(x, width = 30)
    ) +
    scale_y_continuous(labels = scales::percent)
  if (input$rb == TRUE) {
    p <- p +
      geom_point(aes(colour = prop, label = journal_title)) +
      scale_colour_viridis(option = "C") +
      scale_fill_viridis(option = "C") +
      guides(color = FALSE, fill = FALSE)
  } else {
    p <- p
  }
  plotly::ggplotly(p, tooltip = c("label", "y")) %>%
    layout(margin = list(l = 200))
})
```


### Journal hybrid open access share by publisher by year 

```{r}
fillCol(
  height = 600,
  flex = c(NA, 1),
  inputPanel(
    selectInput(
      "Publisher",
      label = "Select Publisher",
      choices = c("All", unique(hybrid_df$publisher)),
      selected = "All"
    )
  ),
  plotlyOutput("yearplot", height = "100%")
)

output$yearplot <- renderPlotly({
  if (input$Publisher == "All") {
    tt <-   hybrid_df %>%
      filter(!is.na(license)) %>%
      group_by(year, publisher, journal_title, yearly_jn_volume) %>%
      summarize(oa = n()) %>%
      mutate(prop = oa / yearly_jn_volume)
  } else {
    tt <-  hybrid_df %>%
      filter(!is.na(license)) %>%
      filter(publisher == input$Publisher) %>%
      group_by(year, publisher, journal_title, yearly_jn_volume) %>%
      summarize(oa = n()) %>%
      mutate(prop = oa / yearly_jn_volume)
  }
  
  p <- ggplot(tt, aes(y = prop, x = as.factor(year))) +
    geom_boxplot() +
    coord_flip() +
    theme_minimal() +
    xlab("Year") +
    ylab(NULL) +
    scale_x_discrete(drop = FALSE) +
    scale_y_continuous(labels = scales::percent)
  if (input$rb == TRUE) {
    p <- p +
      geom_point(aes(colour = prop, label = journal_title)) +
      scale_colour_viridis(option = "C") +
      scale_fill_viridis(option = "C") +
      guides(color = FALSE, fill = FALSE)
  } else {
    p <- p
  }
  plotly::ggplotly(p, tooltip = c("label", "y"))
})
```


Row 
---------------------------------------------

### Table view {data-height=700}

```{r}
hybrid_data <- reactive({
  if (input$Publisher == "All") {
    hybrid_df
  } else {
    hybrid_df <- hybrid_df %>%
      filter(publisher == input$Publisher)
  }
  hybrid_df %>%
    filter(year %in% input$Year) %>%
    group_by(year, publisher, journal_title, yearly_jn_volume) %>%
    summarize(oa = n()) %>%
    mutate(prop = round((oa / yearly_jn_volume) * 100, 2)) %>%
    arrange(desc(year)) %>%
    select(
      Year = year,
      Publisher = publisher,
      Journal = journal_title,
      `Article Volume` = yearly_jn_volume,
      `OA Articles` = oa,
      `OA Share (in %)` = prop
    )
})

renderDataTable({
  hybrid_data()
},
rownames = FALSE,
filter = 'bottom',
options = list(
  pageLength = 10,
  lengthMenu = c(5, 10, 25, 50),
  columnDefs = list(list(
    className = 'dt-head-left', targets = "_all"
  ))
))

# download button
tags$h4("Export to Excel")
downloadHandler(
  filename = function() {
    paste('hybridoa-', Sys.Date(), '.xlsx', sep = '')
  },
  content = function(file) {
    writexl::write_xlsx(hybrid_data(), file)
  }
)
```

Institutional View
=================================================================

```{r}
my_data <- reactive({
  if(!is.null(input$plot_clicked_data$name)) {
    hybrid_df %>%
    filter(suffix == input$plot_clicked_data$name | domain == input$plot_clicked_data$name) %>%
    mutate(plot_title = ifelse(!input$plot_clicked_data$name %in% suffix, 
                               domain, 
                               suffix))
    } else
      hybrid_df
})
```

Column 
----------------------------------------------------------------------

### Text-mined author emails (first match per article) - Double click for selection!  

```{r}
output$plot <- renderEcharts4r({
  hybrid_df %>%
     count(suffix, domain) %>% 
     filter(!is.na(domain)) %>%
     e_charts() %>%
     e_treemap(suffix, domain, n, leafDepth = "1", 
               name = "Corresponding Email") %>%
     e_tooltip(trigger = "item") 
})

output$oa_license <- renderPlotly({
    ggplot(my_data(), aes(factor(year), fill = license)) + 
    geom_bar(position = position_stack(reverse = TRUE)) +
    ggtitle(unique(my_data()$plot_title)) +
    scale_fill_manual(
      "License",
      values =  c(
        "#65BADA",
        "#068E8C",
        "#00A757",
        "#E5BA52",
        "#D86F27",
        "#C82E6B"
      )
    ) +
    scale_x_discrete("", drop = FALSE) +
    scale_y_continuous("Articles",
      labels = function(x)
        format(x, big.mark = " ", scientific = FALSE),
      breaks = scales::pretty_breaks()
    ) +
    theme_minimal() -> p
     tt <- plotly::ggplotly(p, tooltip = c("y"))
  tt$x$data <- lapply(tt$x$data, function(x) {
    x$text <- paste(x$name, x$y, sep = ": ")
    x
  })
  tt
})

output$oa_source <- renderPlotly({
  p <- 
    ggplot(my_data(), aes(factor(year), ..count.., fill = hybrid_type)) +
    geom_bar(position = position_stack(reverse = TRUE)) +
    xlab("Year") +
    ylab("Articles") +
    scale_y_continuous(
      labels = function(x)
        format(x, big.mark = " ", scientific = FALSE),
      breaks = scales::pretty_breaks()
    ) +
    scale_fill_manual(
      "Sources",
      values = c(
        "Open APC (TA)" = "#23445D",
        "Open APC (Hybrid)" = "#AE4132",
        "SCOAP<sup>3</sup>" = "#12AAB5"
      ),
      na.value = "#b3b3b3a0"
    ) +
    theme_minimal()
   tt <- plotly::ggplotly(p, tooltip = c("y"))
  tt$x$data <- lapply(tt$x$data, function(x) {
    x$text <- paste(x$name, x$y, sep = ": ")
    x
  })
  tt
})

fillRow(flex = c(1, 2),
  fillCol(echarts4rOutput("plot")),
  fillCol(plotlyOutput("oa_license"), plotlyOutput("oa_source"))
)
```



Row 
------------------------------------------------------------------- 


### Top Publishers

```{r}
renderDataTable({
 my_data() %>%
      count(publisher) %>%
      arrange(desc(n)) %>%
      mutate(prop = round(n / sum(n) * 100,2))
})
```