Skip to main content

Introduction to Communication and Media Research with R

14 Visualizations

14.1 Introduction

The Imperative of Visual Representations in Research

Data visualization serves as a crucial tool in media and information studies, designed to transform complex data sets into a visually engaging format that enhances comprehension and facilitates insights. The visual representation helps in making sense of large amounts of data, which might be impenetrable in raw form.

In the contemporary landscape of academic and applied research, visual representations have risen to prominence as indispensable components of scholarly communication. Visual aids such as tables, graphs, and charts play a critical role in the dissemination of research findings, serving as a potent mechanism for encapsulating complex datasets, theories, and relationships in an accessible manner (Tufte, 2001). In fact, well-designed visualizations can make a marked difference in how research is consumed, interpreted, and critiqued.

Facilitating Cognitive Processing

Humans are inherently visual creatures. A considerable proportion of the brain is devoted to visual processing, making visual cues one of the most effective means of information transmission (Ware, 2012). When integrated into research papers, visualizations capitalize on this cognitive predisposition, facilitating quicker comprehension and more robust retention of the material presented. This is particularly pertinent in fields like Communication and Media Research Methods, where data are often multifaceted and conclusions nuanced.

Enhancing Analytical Rigor

Visual representations also contribute to the analytical rigor of research endeavors. For instance, a well-plotted graph can illuminate trends and anomalies that may remain obfuscated in raw, numerical data (Cairo, 2016). As such, visualizations can serve as both a diagnostic tool for the researcher and an evidentiary asset for the reader, substantiating claims and underpinning arguments.

Complementing Textual Explanations for a Holistic Understanding

While the textual component of a research paper provides the requisite detail and contextual underpinning, it may fall short in offering the immediacy and clarity that visual aids can deliver.

Symbiosis of Text and Visuals

A synergetic relationship between text and visuals often proves to be the most effective approach for conveying a multi-layered research narrative. Textual explanations can delve into the intricacies, limitations, and theoretical frameworks that underlie the research, while accompanying visual aids provide an immediate, intuitive grasp of key points, trends, and implications (Kelleher & Wagener, 2011).

Accessibility and Engagement

Furthermore, well-crafted visualizations make scholarly work more accessible to a broader audience, including those who may lack specialized training in the subject area. They can break down barriers of jargon and complexity, transforming abstract concepts into tangible insights (Börner, 2015). In an educational context, this dual modality of presentation—textual and visual—appeals to varied learning styles, fostering greater student engagement and comprehension.

In summary, the role of visual representations in research is not merely supplementary; it is integral. Visual aids deepen our understanding of data, enhance analytical rigor, and facilitate clearer, more effective communication of research outcomes.

14.2 Tables

Importance of Tables

Tables are a foundational element in data visualization, offering a structured format for organizing numerical data. By arranging data in rows and columns, tables enable viewers to compare figures systematically and discern patterns or discrepancies with ease.

Concise Representation of Numerical Data

When designing tables for data visualization, it is paramount to arrange data coherently and to label all columns and rows with precision. Such organization ensures that viewers can follow and interpret the data accurately without confusion. Tables stand as one of the most efficient means for the organized presentation of numerical data. By arranging data in rows and columns, tables offer a spatial framework that enables quick scanning and interpretation. This is particularly advantageous in research contexts where large datasets are prevalent, as tables condense this information into a digestible format. Unlike other types of visual aids like graphs and charts, tables can preserve the exact numerical values, which can be crucial for statistical analysis and for the validation of research findings (Zacks & Tversky, 1999).

Guidelines for Creating Tables

Structure and Layout

The utility of a table is directly contingent upon its design and layout. Poorly structured tables can obfuscate rather than clarify, hindering the reader’s comprehension. Hence, attention to the following design elements is crucial:

  • Alignment: Numeric data should be right-aligned for easier comparison, while text is generally left-aligned.
  • Row and Column Spacing: Adequate spacing between rows and columns improves readability.
  • Gridlines: Use subtle gridlines or zebra-striping to delineate rows or columns without overwhelming the visual field (Kosslyn, 2006).

Inclusion of Title, Column Headings, and Footnotes

  • Title: Every table should have a descriptive title that encapsulates the content and purpose of the table. The title should be placed above the table and be clear and concise.

  • Column Headings: These serve as descriptors for the data contained in the respective columns and should be as short as possible while still being descriptive. Headings often contain units of measurement and should be clearly differentiated from the data rows.

  • Footnotes: Utilized for explaining abbreviations, symbols, or methodological points that require clarification, footnotes should be positioned directly below the table and are typically denoted by asterisks or numerical superscripts (American Psychological Association [APA], 2019).

14.3 Illustrations

Types of Illustrations

Diagrams

Illustrations stand out in data visualization by distilling intricate ideas or processes into more digestible, straightforward visual representations. They play a key role in simplifying and clarifying information that may be too complex for text descriptions alone. They are highly useful in research to delineate component parts, explicate functional relationships, or visualize abstract theories. For example, a Venn diagram may be employed to depict the intersections among different sets or groups, effectively elucidating how they relate to one another.

Flowcharts

Flowcharts are specialized diagrams that visually represent a sequence of steps, generally using geometrical shapes connected by lines and arrows to indicate flow direction. They are often employed to describe workflows, algorithms, or processes, allowing for easier understanding and problem-solving. In the context of communication and media research, a flowchart could be used to outline the steps taken in a content analysis procedure, for instance.

Organizational Models

Organizational models provide a structural overview of a system, often portraying hierarchies or relational networks. Such models can be useful in a range of fields, from sociology where they might depict social structures, to business research where they can illustrate organizational hierarchies.

Using Illustrations Effectively

When to Employ Illustrations

The judicious use of illustrations hinges on their appropriateness to the research context. They are most effective when they serve to:

  • Simplify complex processes or systems, aiding in conceptual understanding.
  • Facilitate comparisons or highlight contrasts.
  • Accentuate a focus on specific elements within a broader scope (Carney & Levin, 2002).

While the temptation to include visuals for aesthetic appeal is natural, their primary function should remain informational. Overuse or misuse can divert attention and compromise the integrity of the research (Tufte, 1990).

How to Employ Illustrations for Maximum Impact

  • Clarity and Simplicity: Illustrations should be easily interpretable. Overcomplicated visuals can detract from the information they aim to convey.

  • Labeling: Adequate and accurate labeling provides context and improves comprehension.

  • Scale and Proportion: Ensuring the scale and proportionality of various elements within the illustration is accurate avoids misleading interpretations (Ware, 2004).

  • Color and Texture: While these elements can enhance differentiation, they should be used sparingly to avoid visual clutter.

  • Alignment with Text: It’s essential that the illustrations directly relate to the accompanying text and that cross-references are clear and easy to locate (Few, 2009).

14.4 Images

Incorporating images into data visualizations can significantly enhance the impact of the presented information. They often evoke emotional responses and can make the visualized data more engaging and memorable for the audience. In data visualization, the integration of images must be meticulously considered; they should complement and clarify the data, avoiding distractions or misinterpretations. Relevant images enhance understanding and retain the visualization’s focus.

Types of Images

Photographs

Photographs can be a powerful tool in research, serving as both primary and secondary data. For example, in social research, photographs can help capture societal trends, behaviors, or physical spaces. In media and communication studies, images from media platforms may be used to support qualitative analyses such as content analysis or semiotics.

Note. This is an example of an image you can use while trying to describe a streamer setup when writing about Twitch streamer culture or technology.

Screen Captures

In the digital age, screen captures have become increasingly relevant, especially in fields that study online behaviors, interfaces, and digital content. Researchers might use screen captures to record instances of social media interactions, display website features, or illustrate software functionalities (Mann & Stewart, 2000).

Ethical Considerations

Fair Use in Research Papers

Fair use is a doctrine that permits limited use of copyrighted material without permission for purposes such as criticism, comment, news reporting, teaching, scholarship, and research. However, fair use is a complex legal doctrine and is determined on a case-by-case basis, considering factors like the purpose of use, the nature of the copyrighted work, the amount used in relation to the copyrighted work as a whole, and the effect of the use on the market value of the copyrighted work (United States Copyright Office, 2010).

The ‘fair use’ doctrine can sometimes be invoked in academic research under specific conditions, but it’s advisable to consult legal experts when in doubt. Universities often have resources or legal departments that can offer advice on this (Butler, 2014).

14.5 Plots

Prepare Workspace

Load Libraries

Tidyverse

We load the tidyverse package, which is a collection of R packages designed for data science.

Data.Table

Next, we load the data.table package. It provides an enhanced version of data frames for more efficient data manipulation.

Read in All of Your Data

We use the fread function from the data.table package to read in various datasets from URLs.

Anime Dataset

This dataset presumably contains information related to anime.

anime <- fread("https://raw.githubusercontent.com/rfordatascience/tidytuesday/master/data/2019/2019-04-23/tidy_anime.csv")
Horror Movies Dataset

This dataset likely contains data related to horror movies.

horror_movies <- fread('https://raw.githubusercontent.com/rfordatascience/tidytuesday/master/data/2022/2022-11-01/horror_movies.csv')
Richmond Way Dataset

This dataset could contain information related to Richmond Way, though the exact details would be available in the dataset’s documentation.

richmondway <- fread('https://raw.githubusercontent.com/rfordatascience/tidytuesday/master/data/2023/2023-09-26/richmondway.csv')
Television Ratings Dataset

This dataset likely contains television ratings data.

television <- fread("https://raw.githubusercontent.com/rfordatascience/tidytuesday/master/data/2019/2019-01-08/IMDb_Economist_tv_ratings.csv")
14.5.0.0.1 Video Games Dataset

This dataset presumably contains information about video games.

video_games <- fread("https://raw.githubusercontent.com/rfordatascience/tidytuesday/master/data/2019/2019-07-30/video_games.csv")

Explain the data

Anime

Source: https://github.com/rfordatascience/tidytuesday/blob/master/data/2019/2019-04-23/readme.md

variable class description
animeID double Anime ID (as in https://myanimelist.net/anime/animeID)
name character anime title - extracted from the site.
title_english character title in English (sometimes is different, sometimes is missing)
title_japanese character title in Japanese (if Anime is Chinese or Korean, the title, if available, in the respective language)
title_synonyms character other variants of the title
type character anime type (e.g. TV, Movie, OVA)
source character source of anime (i.e original, manga, game, music, visual novel etc.)
producers character producers
genre character genre
studio character studio
episodes double number of episodes
status character Aired or not aired
airing logical True/False is still airing
start_date double Start date (ymd)
end_date double End date (ymd)
duration character Per episode duration or entire duration, text string
rating character Age rating
score double Score (higher = better)
scored_by double Number of users that scored
rank double Rank - weight according to MyAnimeList formula
popularity double based on how many members/users have the respective anime in their list
members double number members that added this anime in their list
favorites double number members that favorites these in their list
synopsis character long string with anime synopsis
background character long string with production background and other things
premiered character anime premiered on season/year
broadcast character when is (regularly) broadcasted
related character dictionary: related animes, series, games etc.
Horror Movies

Source: https://github.com/rfordatascience/tidytuesday/blob/master/data/2022/2022-11-01/readme.md

Variable Type Definition Example
id int unique movie id 4488
original_title char original movie title Friday the 13th
title char movie title Friday the 13th
original_language char movie language en
overview char movie overview/desc Camp counselors are stalked…
tagline char tagline They were warned…
release_date date release date 1980-05-09
poster_path char image url /HzrPn1gEHWixfMOvOehOTlHROo.jpg
popularity num popularity 58.957
vote_count int total votes 2289
vote_average num average rating 6.4
budget int movie budget 550000
revenue int movie revenue 59754601
runtime int movie runtime (min) 95
status char movie status Released
genre_names char list of genre tags Horror, Thriller
collection num collection id (nullable) 9735
collection_name char collection name (nullable) Friday the 13th Collection
Roy Kent F-ck count

Source: https://github.com/rfordatascience/tidytuesday/blob/master/data/2023/2023-09-26/readme.md#roy-kent-fk-count

variable class description
Character character Character single value - Roy Kent
Episode_order double The order of the episodes from the first to the last
Season double The season 1, 2 or 3 associated with the count
Episode double The episode within the season associated with the count
Season_Episode character Season and episode as a combined variable
F_count_RK double Roy Kent’s F-ck count in that season and episode
F_count_total double Total F-ck count by all characters combined including Roy Kent in that season and episode
cum_rk_season double Roy Kent’s cumulative F-ck count within that season
cum_total_season double Cumulative total F-ck count by all characters combined including Roy Kent within that season
cum_rk_overall double Roy Kent’s cumulative F-ck count across all episodes and seasons until that episode
cum_total_overall double Cumulative total F-ck count by all characters combined including Roy Kent across all episodes and seasons until that episode
F_score double Roy Kent’s F-count divided by the total F-count in the episode
F_perc double F-score as percentage
Dating_flag character Flag of yes or no for whether during the episode Roy Kent was dating the characted Keeley
Coaching_flag character Flag of yes or no for whether during the episode Roy Kent was coaching the team
Imdb_rating double Imdb rating of that episode
TV’s Golden Age

Source: https://github.com/rfordatascience/tidytuesday/blob/master/data/2019/2019-01-08/readme.md

type variable missing complete n min max
character genres 0 2266 2266 5 25
character title 0 2266 2266 1 51
character titleId 0 2266 2266 9 9
Date date 0 2266 2266 1990-01-03 2018-10-10
integer seasonNumber 0 2266 2266 NA NA
numeric av_rating 0 2266 2266 NA NA
numeric share 0 2266 2266 NA NA

Video Games

Source: https://github.com/rfordatascience/tidytuesday/blob/master/data/2019/2019-07-30/readme.md

variable class description
number double Game number
game character Game Title
release_date character Release date
price double US Dollars + Cents
owners character Estimated number of people owning this game.
developer character Group that developed the game
publisher character Group that published the game
average_playtime double Average playtime in minutes
median_playtime double Median playtime in minutes
metascore double Metascore rating

Components of ggplot2 in R

ggplot2 is a data visualization package in R that is part of the tidyverse. This package allows for layering of various graphic components to build complex visualizations.

Basic Syntax

The foundation of any ggplot is the ggplot() function, to which you can add different geoms (geometric objects) to visualize the data.

library(ggplot2)
ggplot(data = data_frame, aes(x = variable1, y = variable2)) + geom_point()

In this line, data_frame is the dataset being visualized, aes() is the function to map variables to aesthetic attributes, and geom_point() adds points to the plot for each combination of x and y values.

Aesthetic Mappings (aes)

The aes() function allows you to map variables in your dataset to aesthetic attributes like x-position, y-position, color, fill, and transparency (alpha).

ggplot(data = data_frame, aes(x = variable1, y = variable2, color = variable3)) + geom_point()

Here, variable3 is mapped to the color aesthetic, resulting in points with colors that reflect the value of variable3.

Labs (Labels)

The labs() function is used to customize or add labels to the ggplot, such as the title and axis labels.

ggplot(data_frame, aes(x = variable1, y = variable2)) + geom_point() + labs(title = "My Plot", x = "X-Axis", y = "Y-Axis")

Pre-Made Themes

ggplot2 comes with several pre-made themes like theme_minimal() and theme_light() that can be easily applied to a plot.

ggplot(data_frame, aes(x = variable1)) + geom_histogram() + theme_light()

Customizing Themes

For more control over the look of your plot, you can use the theme() function and specify various elements.

ggplot(data_frame, aes(x = variable1)) + geom_histogram() + theme(axis.text.x = element_text(angle = 45))

Color Schemes

To set or customize color schemes, you can use scale_color_* and scale_fill_* functions.

ggplot(data_frame, aes(x = variable1, fill = variable2)) + geom_histogram() + scale_fill_brewer(palette = "Blues")

Binwidth

In histograms, the binwidth parameter specifies the width of each bin.

ggplot(data_frame, aes(x = variable1)) + geom_histogram(binwidth = 5)

Legends

Legends in ggplot2 are usually generated automatically but can be customized using the guides() function or directly within scale_* functions.

ggplot(data_frame, aes(x = variable1, color = variable2)) + geom_point() + guides(color = guide_legend("Legend Title"))

This allows you to change the title of the legend from the default to “Legend Title.”

Distribution Plots

This section covers various types of distribution plots including histograms, density plots, violin plots, and boxplots.

Histogram

A histogram is a graphical representation that organizes a group of data points into specified ranges. It is an estimate of the probability distribution of a continuous variable. Histograms are effective in visualizing the frequency distribution of continuous data sets. By grouping data into intervals, histograms provide a clear picture of the distribution’s shape and the prevalence of data points within specific ranges. The data is partitioned into bins, and the number of data points in each bin is represented by the height of the corresponding bar (Wickham, 2016). Here, we are using the richmondway dataset to examine the frequency distribution of “F-ck Count” by the character Roy Kent.

ggplot(richmondway, aes(x = F_count_RK)) +
  geom_histogram() +
  labs(title = "Distribution of Roy Kent's F-ck Count",
       x = "F-ck Count",
       y = "Frequency")
## `stat_bin()` using `bins = 30`. Pick better value with `binwidth`.

Density Plot

Density plots visualize the distribution of a continuous variable over a continuous range. Unlike histograms, these plots are smooth, which makes them suitable for estimating the probability density function of the underlying variable (Silverman, 1986). In this example, we will be using the horror_movies dataset to visualize the density of movie ratings.

Filter for top languages

After identifying which languages to include, create a new data set filtered for these languages.

horror_movies_top_5_languages <- horror_movies %>%
  filter(original_language %in% c("en", "es", "ja", "pt"))
Create Density Plot

After filtering for the top 5 languages, a density plot is created.

ggplot(horror_movies_top_5_languages, aes(x = vote_average)) +
  geom_density(aes(fill = original_language), alpha = 0.5) +
  labs(title = "Density Plot of Horror Movie Ratings",
       x = "Rating",
       y = "Density")

Violin Plot

Violin plots combine features of boxplots and density plots to show the distribution, median, and interquartile range of the data. They are particularly useful for comparing the distributions of multiple categories in a dataset (Hintze & Nelson, 1998). Here, we use the video_games dataset to visualize the average metascores for the top 5 publishers.

Find the 5 Most Prolific Publishers

First, we identify which publishers are most prolific.

video_games %>%
  group_by(publisher) %>%
  count() %>%
  arrange(desc(n)) %>%
  ungroup() %>%
  top_n(5)
## Selecting by n
## # A tibble: 6 × 2
##   publisher          n
##   <chr>          <int>
## 1 Big Fish Games   284
## 2 SEGA             141
## 3 Strategy First   129
## 4 Ubisoft          123
## 5 Square Enix       95
## 6 <NA>              95
Create Violin Plot

After filtering for the top 5 publishers, we generate the violin plot.

video_games_top_5_publishers <- video_games %>%
  filter(publisher %in% c("Big Fish Games", "SEGA", "Strategy First", "Ubisoft", "Square Enix"))

ggplot(video_games_top_5_publishers, aes(x = publisher, y = metascore)) +
  geom_violin() +
  labs(title = "Violin Plot of Average Metascore by Top 5 Publisher",
       x = "Publisher",
       y = "Average Metascore")
## Warning: Removed 572 rows containing non-finite values (`stat_ydensity()`).

Boxplot

A boxplot provides a graphical representation of the central tendency and spread of a dataset, depicting the median, quartiles, and potential outliers. Boxplots are useful for identifying skewness and outliers in the data (Tukey, 1977). We’ll use the anime dataset to explore how scores are distributed across different types of anime.

ggplot(anime, aes(x = type, y = score)) +
  geom_boxplot(fill = "#ff00ff", color = "#770077") +
  labs(title = "Boxplot of Anime Scores by Type",
       x = "Type",
       y = "Score")
## Warning: Removed 174 rows containing non-finite values (`stat_boxplot()`).

Correlation Plots

This section focuses on the use of correlation plots including scatter plots, heatmaps, and bubble plots.

Scatter Plot

Scatter plots are particularly adept at demonstrating the relationship between two quantitative variables. By plotting data points on a two-dimensional graph, they allow for the observation of patterns or correlations within the data. A scatter plot utilizes Cartesian coordinates to display values of two variables, one plotted along the x-axis and the other plotted along the y-axis. It is commonly used to observe and show relationships between two numeric variables (Cleveland, 1994). In this example, we are using the richmondway dataset.

Convert Seasons to Factor

It’s common practice to convert categorical variables to factors when plotting in ggplot2.

richmondway <- richmondway %>%
  mutate(Season = as.factor(Season))
Create Scatter Plot

We then proceed to create the scatter plot.

ggplot(richmondway, aes(x = F_count_total, y = F_count_RK)) +
  geom_point(aes(color = Season), size = 3) +
  theme_minimal() + 
  labs(title = "Roy Kent F-ck Count by IMDB Rating",
       x = "Total F-ck Count",
       y = "F-ck Count")

Connected Scatter Plot

A connected scatter plot combines elements of both scatter and line plots to visualize the relationship between two variables while emphasizing the sequence or progression of data points. Unlike a traditional scatter plot that only displays individual data points, a connected scatter plot links these points with lines, highlighting the order or trend of the data. This method is particularly useful when tracking the progression of two variables in relation to each other over time or categories, and it can reveal patterns that might not be evident in a standard scatter plot. A line plot, on the other hand, typically focuses on showing a trend or change in a single variable over time or categories and is more straightforward in depicting time series data. Line plots are invaluable in data visualization for their ability to illustrate trends and changes over time. The connection of data points with a line makes it straightforward to track progressions or declines within the dataset.

In the following example, we will use the richmondway dataset to create a connected scatter plot. The aim is to plot the average F_count_total for each season, showing how this average changes from one season to the next.

Average F_count_total by Season

First, we need to compute the average F_count_total for each season. This will ensure that each season is represented by a single data point in the plot.

richmondway_avg <- richmondway %>%
  group_by(Season) %>%
  summarize(Avg_F_count = mean(F_count_total))
Create Connected Scatter Plot

Now, we create the connected scatter plot using ggplot2. This plot will show the average F_count_total for each season and connect these averages to illustrate the trend over the seasons.

ggplot(richmondway_avg, aes(x = Season, y = Avg_F_count)) +
  geom_point(aes(color = Season), size = 3) +
  geom_line(aes(group = 1), color = "blue") +
  theme_minimal() +
  labs(title = "Average F-ck Count Across Seasons",
       x = "Season",
       y = "Average Total F-ck Count")

In this connected scatter plot, each point represents the average total count of a particular term (in this case, “F-ck Count”) for a season. The lines connecting these points demonstrate the progression or trend of this average over the seasons, providing a clear visualization of how the average count changes from one season to the next. The use of different colors for each season can further aid in distinguishing the data points, enhancing the overall understanding of the trends displayed.

Heatmap

A heatmap is a data visualization technique that represents the magnitude of observations as color in a two-dimensional plane. This plot is often used to understand complex data structures and correlations between multiple variables (Wilkinson & Friendly, 2009). In this example, we use the horror_movies dataset.

correlation_matrix <- cor(horror_movies[,c("popularity", "vote_count", "vote_average", "budget", "revenue")], use = "complete.obs")
ggplot(melt(correlation_matrix), aes(x=Var1, y=Var2)) +
  geom_tile(aes(fill=value), colour="white") +
  scale_fill_gradient(low="white", high="blue")
## Warning in melt(correlation_matrix): The melt generic in data.table has been
## passed a matrix and will attempt to redirect to the relevant reshape2 method;
## please note that reshape2 is deprecated, and this redirection is now deprecated
## as well. To continue using melt methods from reshape2 while both libraries are
## attached, e.g. melt.list, you can prepend the namespace like
## reshape2::melt(correlation_matrix). In the next version, this warning will
## become an error.

Bubble Plot

A bubble plot is an extension of the scatter plot, where a third dimension of data is added through the size of the bubbles. This allows for the simultaneous comparison of three variables (Cleveland, 1994). We use the video_games dataset in this example.

14.5.0.0.2 Remove Free Titles and Titles with Missing Data

First, we remove the rows with missing values and rows where any column has a 0 value.

nonzero_video_games <- video_games %>%
  filter(complete.cases(.)) %>%  # Remove rows with missing values (NA)
  filter(!rowSums(. == 0))  # Remove rows where any column has a 0 value
Create Bubble Plot

We then proceed to create the bubble plot.

ggplot(nonzero_video_games, aes(x = median_playtime, y = metascore, size = price)) +
  geom_point(aes(color = owners), alpha = 0.6) +
  labs(title = "Bubble Plot of Playtime, Metascore, Price, and Ownership",
       x = "Median Playtime",
       y = "Metascore")

Ranking Plots

This section will cover the creation of ranking plots, specifically bar plots and lollipop plots.

Bar Plot

Bar plots represent categorical data with rectangular bars where the lengths are proportional to the counts or values they represent. Bar plots can be oriented horizontally or vertically and are useful for comparing quantities across categories (Wickham, 2016). Bar plots are preferred over line plots when the goal is to compare discrete categories or distinct groups. The segmented nature of bar plots makes them ideal for highlighting differences between items without implying a continuous sequence.

Identify Top Anime Producers

In this code snippet, we identify the top anime producers in the anime dataset.

anime %>%
  group_by(producers) %>%
  count() %>%
  arrange(desc(n)) %>%
  ungroup() %>%
  top_n(6)
## Selecting by n
## # A tibble: 6 × 2
##   producers           n
##   <chr>           <int>
## 1 ""              16559
## 2 "TV Tokyo"       1940
## 3 "Aniplex"        1721
## 4 "Bandai Visual"  1678
## 5 "Lantis"         1630
## 6 "Movic"          1288
Create a New Data Frame for Just the Top Five Producers

Here we filter the anime data to only include records from the top five producers.

anime_top_5_producers <- anime %>%
  filter(producers %in% c("TV Tokyo", "Aniplex", "Bandai Visual", "Lantis", "Movic"))
Create New Data Frame with the Average Score by Producer

We then calculate the average score by producer. Note the use of na.rm = TRUE to remove NA values for an accurate mean.

anime_avg_score_by_producer <- anime_top_5_producers %>%
  group_by(producers) %>%
  summarise(avg_score = mean(score, na.rm = TRUE))  # Remove NA values for accurate mean
Create Bar Plots

Finally, a bar plot is created using the average scores by producer.

ggplot(anime_avg_score_by_producer, aes(x = reorder(producers, -avg_score), y = avg_score)) +
  geom_bar(stat = "identity") +
  labs(title = "Bar Plot of Average Score by Producer",
       x = "Producers",
       y = "Average Score")

Lollipop Plot

A lollipop plot combines elements of bar plots and scatter plots to represent values of different categories. It consists of a stem (akin to the bar in a bar plot) and a dot (akin to the point in a scatter plot) at the end of the stem to signify the value (Tufte, 2001).

Separate Genres into New Rows

Assuming television is your original dataset, this code separates each genre into a new row.

# Assuming `television` is your original data frame
long_television <- television %>%
  separate_rows(genres, sep = ",\\s*")  # Separate by comma and any subsequent whitespace
Identify Top Genres

This code identifies the top 10 genres from the television dataset.

television_top_genres_list <- long_television %>%
  group_by(genres) %>%
  count() %>%
  arrange(desc(n)) %>%
  ungroup() %>%
  top_n(10)
## Selecting by n
television_top_genres_list
## # A tibble: 10 × 2
##    genres        n
##    <chr>     <int>
##  1 Drama      2266
##  2 Crime       822
##  3 Mystery     558
##  4 Comedy      516
##  5 Action      387
##  6 Romance     235
##  7 Fantasy     223
##  8 Adventure   204
##  9 Thriller    160
## 10 Sci-Fi      154
Create New Data Frame for Just the Top 10 Genres

Here we filter the data to only include records from the top 10 genres.

television_top_genres <- long_television %>%
  filter(genres %in% television_top_genres_list$genres)
Calculate Average Rating for Each Genre

This code calculates the average rating for each genre.

television_top_genres_avg_share <- television_top_genres %>%
  group_by(genres) %>%
  summarise(share = median(share, na.rm = TRUE))
Create Lollipop Plot

Finally, a lollipop plot is created to visualize the average rating for each of the top 10 genres.

ggplot(television_top_genres_avg_share, aes(x = reorder(genres, -share), y = share)) +
  geom_segment(aes(xend = genres, yend = 0), color = "blue") +
  geom_point(size = 3, color = "red") +
  labs(title = "Lollipop Plot of Average TV Show Share by Genre",
       x = "TV Show Genre",
       y = "Average Share")

Part of a Whole

This section will cover various methods for depicting part-of-a-whole relationships in data visualization, including pie charts, grouped and stacked bar plots, and treemaps.

Pie Chart

A pie chart is a circular chart that divides data into slices to illustrate numerical proportion. Each slice represents a part-to-whole relationship (Bertin, 1983).

14.5.0.0.3 Dataset: Roy Kent F-ck count

A pie chart is created to visualize Roy Kent’s usage of the word “F-ck” across different seasons.

ggplot(richmondway, aes(x = "", y = F_count_RK, fill = factor(Season))) +
  geom_bar(width = 1, stat = "identity") +
  coord_polar("y") +
  labs(title = "Pie Chart of Roy Kent's F-ck Count by Season")

Grouped + Stacked Barplot

Reshape Data

The data is reshaped to facilitate the creation of grouped and stacked bar plots.

television_grouped <- television_top_genres %>%
  group_by(genres) %>%
  summarise(av_rating = mean(av_rating, na.rm = TRUE),
            share = mean(share, na.rm = TRUE)) %>%
  pivot_longer(cols = c(av_rating, share), names_to = "metric")
Create Grouped Barplot

A grouped barplot represents categorical data with multiple sub-categories. It uses adjacent bars to represent the different sub-categories within each primary category, facilitating direct comparison (Wickham, 2016). A grouped bar plot is created to show both average rating and share for each genre of television show.

ggplot(television_grouped, aes(x = genres, y = value, fill = metric)) +
  geom_bar(stat = "identity", position = "dodge") +
  labs(title="Television Shows: Average Rating and Share by Genre", y="Value") +
  theme(axis.text.x = element_text(angle = 45, hjust = 1))
Create Stacked Barplot

A stacked barplot is similar to a standard bar plot but divides each bar into multiple sub-categories. This allows for the representation of part-to-whole relationships within each category (Wickham, 2016). A stacked bar plot is created to show both popularity and rating of horror movies by original language.

ggplot(television_grouped, aes(x = genres, y = value, fill = metric)) +
  geom_bar(stat = "identity", position = "stack") +
  labs(title="Television Shows: Average Rating and Share by Genre", y="Value")

Treemap

Dataset: Horror Movies

A treemap displays hierarchical data as a set of nested rectangles. Each level in the hierarchy is represented by a colored rectangle (‘branch’), which is then sub-divided into smaller rectangles (‘leaves’) (Shneiderman, 1992). A treemap is created to visualize the vote count by original language for horror movies.

library(treemap)

treemap(horror_movies, 
        index = "original_language", 
        vSize = "vote_count", 
        title="Horror Movies: Vote Count by Language")

Store and Edit Plots

Filter data for unique show titles

tv_top_genres_list <- long_television %>%
  distinct(title, genres, .keep_all = TRUE) %>%
  group_by(genres) %>%
  count() %>%
  arrange(desc(n)) %>%
  ungroup() %>%
  top_n(10, wt = n)

tv_top_genres_list
## # A tibble: 10 × 2
##    genres        n
##    <chr>     <int>
##  1 Drama       868
##  2 Crime       272
##  3 Mystery     180
##  4 Comedy      177
##  5 Action      146
##  6 Fantasy      90
##  7 Romance      82
##  8 Sci-Fi       76
##  9 Thriller     74
## 10 Adventure    72

You can store a plot into your environment for easier recall.

tv_genres_bar <- ggplot(tv_top_genres_list, aes(x = reorder(genres, -n), y = n)) +
  geom_bar(stat = "identity") +
  labs(title = "Bar Plot of Show Count by Major Genre",
       x = "Genres",
       y = "Number of shows")

A stored plot can be added to without having to type out all of the code if it is previously stored in the environment.

First load package with unique themes

Add to your stored plot

tv_genres_bar +
  theme_wsj()

Saving Plots in R Markdown

Once you’ve generated and refined your plot, it’s crucial to understand how to save it for further use, whether for presentation, publication, or collaboration. R Markdown, together with the ggplot2 library, offers flexible ways to save your plots in various formats.

Saving Plots Directly in R Markdown

Store plot you want to save locally.

tv_genres_bar_wsj <- tv_genres_bar +
  theme_wsj()

The above chunk will save the plot in the specified directory with the name “plot_name” followed by a figure number (e.g., “plot_name1.png”).

Saving Plots Using ggsave()

The ggsave() function, which comes with the ggplot2 library, provides a straightforward way to save the last plot that you created. Below are the steps and explanations:

  1. PNG Format: Suitable for web applications.
ggsave("tv_genres_bar.png", plot = tv_genres_bar_wsj, width = 10, height = 5)
  1. JPEG Format: Generally used for photographs on the web, but it’s lossy, meaning some image quality is compromised.
ggsave("tv_genres_bar.jpg", plot = tv_genres_bar_wsj, width = 10, height = 5)
  1. PDF Format: Perfect for publications and where high quality is paramount. It retains the quality regardless of how much you zoom.
ggsave("tv_genres_bar.pdf", plot = tv_genres_bar_wsj, width = 10, height = 5)
  1. SVG Format: A vector format suitable for web applications where scalability without loss of resolution is important.
library(svglite)

ggsave("tv_genres_bar.svg", plot = tv_genres_bar_wsj, width = 10, height = 5)

Explanation

  • filename: The name you want to give to the saved plot, which also specifies the format based on the extension.
  • plot: The specific plot you want to save. In this case, it’s tv_genres_bar + theme_wsj().
  • width and height: The width and height of the saved plot in inches.

References

  • Bertin, J. (1983). Semiology of Graphics: Diagrams, Networks, Maps.
  • Cleveland, W. S. (1994). The Elements of Graphing Data.
  • Hintze, J. L., & Nelson, R. D. (1998). Violin Plots: A Box Plot-Density Trace Synergism.
  • Shneiderman, B. (1992). Tree visualization with tree-maps: 2-d space-filling approach. ACM Transactions on Graphics, 11(1), 92–99.
  • Silverman, B. W. (1986). Density Estimation for Statistics and Data Analysis.
  • Tufte, E. R. (2001). The Visual Display of Quantitative Information.
  • Tukey, J. W. (1977). Exploratory Data Analysis.
  • Wickham, H. (2016). ggplot2: Elegant Graphics for Data Analysis. Springer-Verlag New York.
  • Wilkinson, L., & Friendly, M. (2009). The History of the Cluster Heat Map. American Statistician, 63(2), 179–184.