Purpose of the Project

Air transportation is one of the most widely used forms of travel in the United States and around the world. On a typical day, the Federal Aviation Administration (FAA) oversees an average of 44,360 flights in the United States, and during peak periods more than 5,000 aircraft can be in the air at the same time. Whether traveling for business, vacation, family commitments, or waiting for a package to arrive, many Americans depend on air transportation upon a regular basis. My own family provides an example of this, as my mother averages more than fifty flights per year for work.

Boeing and Airbus are the duopoly that dominates the skies. Together, the two companies control more than 80% of the active commercial aircraft fleet worldwide. That said, whenever someone boards a commercial aircraft, there is a good chance the plane was built by either Boeing or Airbus. Because air travel plays such an important role in daily life, it is important for the public to understand the companies that make modern commercial aviation possible and the news that surrounds them. 

The purpose of this project was to analyze and compare media coverage of Boeing and Airbus between September 2021 and November 2026. Using data collected through Media Cloud, the project examined overall coverage volume and identified the most common words and phrases associated with each company in news headlines. By comparing coverage patterns, headline language, and recurring themes, the project provides insight into how Boeing and Airbus were portrayed in the media during that time.

Data Source

Media coverage data for this project were obtained through Media Cloud, an open-source platform that collects and analyzes news content from thousands of media outlets around the world. Media Cloud allows users to search for topics, organizations, and keywords and then retrieve information about news coverage volume, article metadata, and headlines over a selected time period.

For this project, Media Cloud was used to collect four datasets related to The Boeing Company and Airbus: Boeing Story Counts, Airbus Story Counts, Boeing Headlines, and Airbus Headlines. The story count datasets contain daily counts of news articles matching both Boeing and Airbus related search queries, while the headline datasets contain article headlines, publication dates, media source information, and other metadata associated with each article.

The data for this project was collected from the period between September 2021 and November 2026. The datasets were exported from Media Cloud in comma-separated value (.csv) format and were uploaded to a GitHub repository that could be accessed directly within R using raw GitHub URLs. This approach allows the project data to be stored, shared, and reproduced without requiring users to manually download the files before running the code.

Output

The project produced five primary visualizations designed to compare media coverage of Boeing and Airbus. The first visualization is an interactive Plotly line chart displaying weekly story count volume over time. This chart allows users to examine changes in coverage volume for both companies between September 2021 and November 2026. 

The second and third visualizations are unigram frequency charts created using ggplot2. These charts identify the top 10 most common individual words appearing in both Boeing and Airbus related news headlines. Unigram analysis was used to identify top themes and topics most frequently associated with each manufacturer in media coverage. 

The fourth and fifth visualizations are n-gram frequency charts, also created using ggplot2. These charts identify the top 10 most common two-word phrases appearing in Boeing and Airbus related news headlines. While the unigram analysis highlights individual terms, the n-gram analysis provides additional context by showing recurring phrases and topic combinations associated with each company. Two-word phrases were selected because they provide more context than individual words while remaining concise and easy to interpret. For example, Boeing unigram terms such as “737” and “Max” have limited meaning when viewed separately. However, the n-gram phrase “737 Max” provides significantly more context and allows viewers to better understand the topic being discussed. Longer phrase lengths, such as three, four, or five word phrases, appeared less frequently and produced results that were more fragmented and difficult to compare across companies. 

The project uses an interactive Plotly visualization for coverage volume because users can explore changes in coverage across a five-year period by hovering over individual data points. The remaining visualizations were created using ggplot2 because the information they present can be effectively communicated through static charts. Interactive features would add little value to these visualizations because the word and phrase frequencies are already displayed directly on the charts. 

These visualizations were selected because they make it easy for viewers to identify and compare the major themes associated with Boeing and Airbus. Together, the charts provide insight into both the volume of media attention received by each company and the topics most frequently connected to them in news coverage.

Role in Developing The Project

While Microsoft Copilot assisted with coding support, troubleshooting, and script development, I served as the project’s designer and analyst. Below are the main tasks I did in development.

  • Defined the project topic. I selected Boeing and Airbus as the subjects of analysis because they dominate the global commercial aviation industry and play a significant role in both personal travel and global commerce. Because these two companies manufacture most of the world’s commercial aircraft, I believed it would provide an interesting study to look at how the two major aviation companies are represented in the news. 

  • Identified and collected the data. I used Media Cloud to gather story count and headline datasets for both Boeing and Airbus between September 2021 and November 2026. I selected Media Cloud not only because I was already familiar with the platform, but also because it provides access to a large collection of news content from numerous media outlets and allows both coverage volume and headline language to be analyzed over long periods of time. 

  • Organized and managed the datasets. After collecting the Media Cloud exports, I organized the datasets into separate story count and headline files for each company. I then created a GitHub repository to store the data and generated raw GitHub URLs that allowed R to retrieve the files directly. Initially, I planned to use local files stored on my computer because that was how I originally ran the visualizations. It was not until later in the project that I realized I needed a way to share the data so that others could reproduce the analysis. GitHub ultimately turned out to be both easier and more practical because it allowed the project to run directly from online data sources rather than relying on local files stored on my computer. 

  • Developed the analytical approach. After reviewing the available data, I decided to examine Boeing and Airbus from two different perspectives. Story count data would measure the amount of media attention each company received, while headline data would reveal the specific themes and topics most closely associated with each manufacturer. This approach allowed the project to analyze both the quantity and content of news coverage. 

  • Selected the project’s visualizations. I chose to create an interactive Plotly chart to display coverage volume trends over time and used ggplot2 to create unigram and n-gram visualizations. Because the project covers more than five years of news coverage, I felt an interactive chart would make the data easier to explore. Rather than forcing users to estimate values from a crowded graph, the Plotly visualization allows them to hover over the lines and see exact figures. I chose static ggplot2 charts for the unigram and n-gram analyses because those visualizations communicate frequency information clearly and do not benefit significantly from interactivity. 

  • Determined how headline language would be analyzed. I selected unigram analysis to identify the most common individual words appearing in headlines and n-gram analysis to identify commonly occurring phrases. I ultimately chose two-word phrases because they provided additional context while remaining easy to interpret and compare across companies. As mentioned above through unigram analysis, two individual words may not hold significance such as “737” and “Max”. Together though, though n-gram analysis, “737 Max” immediately identifies a specific aircraft and topic.

  • Evaluated and refined the results. After reviewing the analysis output, I removed stop words and other non-informative terms that added little analytical value. Examples included words such as “says,” “ap,” “us,” “new,” “box,” and “de.” I applied the same filtering decisions across both Boeing and Airbus analyses to keep the comparison consistent. I also spent considerable time refining chart colors, labels, titles, sizing, and overall formatting to improve readability and presentation. One of the more significant formatting challenges involved chart dimensions and layout, which required innumerable rounds of adjustments before I was finally able to adjust the appearance. I discuss that challenging experience in greater detail in the “Things I Learned” section.

  • Used AI as a coding assistant. Microsoft Copilot assisted with generating, modifying R code, troubleshooting errors, and suggesting approaches for creating visualizations. However, decisions regarding the project’s topic, data collection strategy, GitHub repository setup, analytical methods, stop-word selection, visualization design, interpretation of results, and final presentation were made by me. In many cases, AI-generated code required modification before it produced the results I wanted, particularly when formatting charts and preparing the final published R Markdown documents. 

Things I learned

I learned a number of things while developing this project, ranging from technical skills to a better understanding of how media coverage shapes public perception of major companies. 

One of the most practical lessons involved learning how to use GitHub as a data repository. Prior to this project, I primarily worked with files given to me and stored locally on my computer. I initially planned to do the same with this project and did not intend to use GitHub until I realized “how am I going to share this raw data with others?” Learning how to upload datasets to GitHub, generate raw GitHub URLs, and import data directly into R made the project more reproducible and ultimately easier to manage. By the end of the project, I was no longer using locally stored datasets such as “boeing_headlines” and “airbus_headlines” files. Instead, I was importing the data directly from GitHub through the corresponding raw URLs.

I also gained a better understanding of text-mining techniques, particularly unigram and n-gram analysis. Before completing this project, I understood the general concepts behind unigrams, bigrams, and trigrams, but I never really considered the strategic advantages and disadvantages of each approach when trying to understand a broader topic. I knew from the beginning that I wanted to use unigrams to identify common terms, but I struggled with whether to use bigram or trigram analysis to better understand the context surrounding those terms. Ultimately, I chose bigrams because they provided additional context while remaining easy to interpret. Singular words “plane” and “crash” do not tell viewers much on their own. However, the phrase “plane crash” grabs attention, provides more context, and makes it much easier to understand the topic being discussed. While trigrams can provide even more context, they appeared less frequently in the data and often produced results that were more fragmented and difficult to compare across companies. 

Another thing I learned was the importance of visualization design. I discovered that producing accurate results is only part of the process. Presenting those results clearly can be equally challenging, as I mentioned in my “Role In Developing The Project” section. I spent a significant amount of time adjusting chart formatting and overall layout within the final R Markdown document. One of the most frustrating challenges involved chart dimensions. When viewed in RStudio, the charts appeared much larger than I wanted, which can be seen in my first iteration of the project. I initially assumed the solution would require modifications to the visualization code to make the charts appear smaller, as shown in the final version. Unfortunately, that was not the case. I eventually learned that chart dimensions are controlled within the R Markdown document, not the R script itself. After significant time troubleshooting with AI assistance, I eventually discovered that the solution was simply adjusting the chunk settings in the R Markdown document to {r, echo=FALSE, fig.width=x, fig.height=x}. What made this issue particularly frustrating was that AI initially convinced me that changing the chunk settings was not the answer. After trying several other solutions, I eventually tested it on my own and discovered it was the correct fix all along. The answer seems simple in hindsight, but learning that chart sizing occurs within the R Markdown document rather than the visualization code itself was one of the more frustrating lessons I learned during the project. 

Finally, the project expanded my understanding of how to work with AI as a coding assistant. While AI was helpful for generating code, explaining functions, and troubleshooting errors, I learned that successful use of AI still requires human oversight, decision-making, and using the “thing between your ears.” Many of the project’s most important decisions, including data collection, visualization selection, stop-word removal, chart formatting, and interpretation of results, required my own judgment. The experience taught me that AI can be a valuable development tool, but it works best when paired with an understanding of what the final product should accomplish and a willingness to question, test, and refine its suggestions. 

So What Do I Think

Overall, the visualizations suggest that Boeing received substantially more media coverage than Airbus during the study period. The headline and phrase analyses indicate that much of Boeing’s coverage centered on aircraft accidents, safety concerns, legal issues, government affairs, and financial performance. These themes are reflected in common words and phrases such as “737 MAX,” “plane crash,” and “Air Force,” which were frequently associated with major news events involving the company. 

Several major news events appear to correspond with the large spikes in Boeing’s coverage volume. One prominent example occurred in January 2024, when a door plug detached midair from an Alaska Airlines Boeing 737 MAX shortly after takeoff, forcing an emergency landing and renewing public and regulatory scrutiny of the aircraft. Another significant spike appears to be associated with Boeing’s legal settlement regarding allegations that the company misled regulators about the 737 MAX prior to two fatal crashes in 2018 and 2019 that collectively killed 346 people and led to the worldwide grounding of the aircraft. The phrase “Air Force” also emerged as one of the most common Boeing related bigrams, largely due to extensive coverage surrounding a Boeing 747-8 aircraft gifted by Qatar for use as Americas new Air Force One Plane. Additionally, some coverage was associated with the crash of an Air India Boeing 787 Dreamliner, an event that drew considerable media attention because it was widely reported as the first fatal crash involving that aircraft model. 

In comparison, Airbus generally received less media attention throughout the period and experienced fewer dramatic coverage spikes. The Airbus visualizations suggest that coverage was more frequently associated with commercial aviation operations, airline business activity, aircraft orders, and broader industry developments rather than major safety incidents or legal controversies. While some Airbus coverage involved individual incidents, these topics appeared less prominent overall than the crisis-related themes observed in Boeing coverage. For example, one of the most common Airbus bigrams was “emergency landing,” which was associated with several separate incidents involving cabin smoke, technical issues, and precautionary landings rather than a single major event. Another notable finding was the appearance of “Air India” among the most common Airbus bigrams. Although the highly publicized Air India crash involved a Boeing aircraft rather than an Airbus, its presence in Airbus coverage highlights how media discussions can overlap across manufacturers. Coverage often compares the competing aircraft manufacturers or discusses broader industry news, resulting in certain companies appearing in stories that are not directly related to their own aircraft.

Overall, this project left me with one final thought. A common phrase often associated with Boeing Company during its dominance in the 1980s and 1990s was, “If it ain’t Boeing, I ain’t going.” As a consumer today, my conclusion is a little different: “If it’s Boeing, I’m definitely not going.”

How the Code Words

Here is a summary, written with assistance from Copilot and reviewed by me, of what each step in the script accomplishes. 

  • STEP 0 - Installs and loads the required R packages needed for data management, text mining, date processing, visualization, and interactive graphics 
  • STEP 1 - Imports Boeing and Airbus headline and story count datasets from GitHub using raw GitHub URLs, allowing the project to access Media Cloud data without requiring local copies of the datasets. 
  • STEP 2 - Combines Boeing and Airbus story count data, calculates weekly coverage totals, and creates an interactive Plotly line chart comparing media coverage volume for both companies over time. 
  • STEP 3 - Converts Boeing news headlines into individual words using unigram analysis, removes stop words and media-related artifacts, filters custom stop words, identifies the most common words appearing in Boeing headlines, and creates a frequency chart displaying the top Boeing headline terms. 
  • STEP 4 - Performs the same unigram analysis process on Airbus headlines, identifies the most common words appearing in Airbus-related coverage, and creates a frequency chart displaying the top Airbus headline terms. 
  • Step 5 - Converts Boeing headlines into two-word phrases using n-gram analysis, removes phrases containing excessive stop words, filters custom phrases, identifies the most common recurring phrases appearing in Boeing headlines, and creates a frequency chart displaying the most common Boeing headline phrases. 
  • STEP 6 - Performs the same two-word phrase analysis on Airbus headlines, identifies the most common recurring phrases appearing in Airbus-related coverage, and creates a frequency chart displaying the most common Airbus headline phrases. 
  • STEP 7 - Displays the completed Plotly and ggplot visualizations generated by the script, including the coverage volume chart, Boeing and Airbus unigram charts, and Boeing and Airbus n-gram charts.  

In short, the script imports Boeing and Airbus news data collected through Media Cloud, compares the amount of media attention received by each company over time, analyzes the words and phrases most commonly associated with each manufacturer in news headlines, and produces both interactive and static visualizations that help viewers compare how Boeing and Airbus are represented in news coverage. 

############################################################
# Boeing vs Airbus Media Coverage Analysis
############################################################

############################################################
# Step 0: Install and Load Packages
############################################################

library(tidyverse)
library(tidytext)
library(lubridate)
library(plotly)

############################################################
# Step 1: Import Data from GitHub
############################################################

airbus_headlines_url <- "https://raw.githubusercontent.com/Deyonnalan/Boeing_Airbus_Data/refs/heads/main/Airbus%20Headlines.csv"

airbus_counts_url <- "https://raw.githubusercontent.com/Deyonnalan/Boeing_Airbus_Data/refs/heads/main/Airbus%20Story%20Counts.csv"

boeing_headlines_url <- "https://raw.githubusercontent.com/Deyonnalan/Boeing_Airbus_Data/refs/heads/main/Boeing%20Headlines.csv"

boeing_counts_url <- "https://raw.githubusercontent.com/Deyonnalan/Boeing_Airbus_Data/refs/heads/main/Boeing%20Story%20Counts.csv"

airbus_headlines <- read_csv(airbus_headlines_url)
airbus_counts <- read_csv(airbus_counts_url)

boeing_headlines <- read_csv(boeing_headlines_url)
boeing_counts <- read_csv(boeing_counts_url)

############################################################
# Step 2: Coverage Volume Analysis
############################################################

boeing_plot <- boeing_counts %>%
  select(date, count) %>%
  mutate(company = "Boeing")

airbus_plot <- airbus_counts %>%
  select(date, count) %>%
  mutate(company = "Airbus")

coverage_data <- bind_rows(
  boeing_plot,
  airbus_plot
)

weekly_counts <-
  coverage_data %>%
  mutate(
    week = floor_date(
      date,
      unit = "week"
    )
  ) %>%
  group_by(
    company,
    week
  ) %>%
  summarise(
    stories = sum(count),
    .groups = "drop"
  )

weekly_plotly <- plot_ly(
  data = weekly_counts,
  x = ~week,
  y = ~stories,
  color = ~company,
  colors = "Set1",
  type = "scatter",
  mode = "lines+markers",
  marker = list(size = 5),
  line = list(width = 2),
  customdata = ~format(stories, big.mark = ","),
  hovertemplate = paste(
    "<b>Company:</b> %{fullData.name}<br>",
    "<b>Week:</b> %{x}<br>",
    "<b>Stories:</b> %{customdata}<extra></extra>"
  )
)

weekly_plotly <- layout(
  weekly_plotly,
  title = "Weekly Media Coverage Volume",
  xaxis = list(
    title = "Week"
  ),
  yaxis = list(
    title = "Number of Stories"
  ),
  legend = list(
    title = list(
      text = "Company"
    )
  ),
  hovermode = "x unified"
)

############################################################
# Step 3: Boeing Unigram Analysis
############################################################

############################################################
# 3A: Create Boeing Headline Unigrams
############################################################

boeing_unigrams <- boeing_headlines %>%
  unnest_tokens(
    word,
    title
  )

############################################################
# 3B: Load and Remove Stop Words
############################################################

data(stop_words)

boeing_unigrams_clean <- boeing_unigrams %>%
  anti_join(
    stop_words,
    by = "word"
  )

############################################################
# 3C: Remove Web and Media Artifacts
############################################################

media_noise <- c(
  "https",
  "http",
  "www",
  "com",
  "amp",
  "rt",
  "co",
  "tco"
)

boeing_unigrams_clean <- boeing_unigrams_clean %>%
  filter(
    !word %in% media_noise
  )

############################################################
# 3D: Examine Most Common Words
############################################################

boeing_unigram_counts <- boeing_unigrams_clean %>%
  count(
    word,
    sort = TRUE
  )

head(
  boeing_unigram_counts,
  30
)

############################################################
# Step 4: Airbus Unigram Analysis
############################################################

############################################################
# 4A: Create Airbus Headline Unigrams
############################################################

airbus_unigrams <- airbus_headlines %>%
  unnest_tokens(
    word,
    title
  )

############################################################
# 4B: Load and Remove Stop Words
############################################################

data(stop_words)

airbus_unigrams_clean <- airbus_unigrams %>%
  anti_join(
    stop_words,
    by = "word"
  )

############################################################
# 4C: Remove Web and Media Artifacts
############################################################

media_noise <- c(
  "https",
  "http",
  "www",
  "com",
  "amp",
  "rt",
  "co",
  "tco"
)

airbus_unigrams_clean <- airbus_unigrams_clean %>%
  filter(
    !word %in% media_noise
  )

############################################################
# 4D: Examine Most Common Words
############################################################

airbus_unigram_counts <- airbus_unigrams_clean %>%
  count(
    word,
    sort = TRUE
  )

head(
  airbus_unigram_counts,
  30
)

############################################################
# 4E: Remove Custom Stop Words
############################################################

custom_stop_words <- tibble(
  word = c(
    "boeing",
    "airbus",
    "says",
    "said",
    "de",
    "news",
    "box",
    "new",
    "ap",
    "us",
    "u.s"
  )
)

airbus_unigrams_final <- airbus_unigrams_clean %>%
  anti_join(
    custom_stop_words,
    by = "word"
  )

############################################################
# 4F: Count Final Unigrams
############################################################

airbus_unigram_counts_final <- airbus_unigrams_final %>%
  count(
    word,
    sort = TRUE
  )

############################################################
# 4G: Select Top Words
############################################################

airbus_top_words <- airbus_unigram_counts_final %>%
  slice_max(
    n,
    n = 10
  )

############################################################
# 4H: Create Visualization
############################################################

airbus_unigram_chart <- ggplot(
  airbus_top_words,
  aes(
    x = reorder(word, n),
    y = n
  )
) +
  geom_col(fill = "firebrick") +
  geom_text(
    aes(label = n),
    hjust = -0.1,
    size = 3
  ) +
  coord_flip() +
  scale_y_continuous(
    expand = expansion(mult = c(0, .10))
  ) +
  labs(
    title = "10 Most Common Words in Airbus Headlines",
    x = "",
    y = ""
  ) +
  theme_minimal()

airbus_unigram_chart
############################################################
# Step 5: Boeing N-Gram Analysis
############################################################

############################################################
# 5A: Choose Phrase Length
############################################################

n_size <- 2

############################################################
# 5B: Load Stop Words
############################################################

data(stop_words)

############################################################
# 5C: Create Boeing N-Grams
############################################################

boeing_ngrams <- boeing_headlines %>%
  unnest_tokens(
    ngram,
    title,
    token = "ngrams",
    n = n_size
  )

############################################################
# 5D: Calculate Stop Word Percentage
############################################################

max_stopword_pct <- 33

boeing_ngrams_filtered <- boeing_ngrams %>%
  mutate(
    words = str_split(
      ngram,
      " "
    )
  ) %>%
  rowwise() %>%
  mutate(
    total_words = length(words),
    stop_words_in_phrase =
      sum(words %in% stop_words$word),
    stop_word_pct =
      stop_words_in_phrase /
      total_words * 100
  ) %>%
  ungroup() %>%
  filter(
    stop_word_pct < max_stopword_pct
  )

############################################################
# 5E: Count Phrases
############################################################

boeing_ngram_counts <- boeing_ngrams_filtered %>%
  count(
    ngram,
    sort = TRUE
  )

############################################################
# 5F: Remove Custom Phrases
############################################################

custom_phrases <- tibble(
  ngram = c(
    "ap top"
  )
)

boeing_ngram_counts_final <- boeing_ngram_counts %>%
  anti_join(
    custom_phrases,
    by = "ngram"
  )

############################################################
# 5G: Select Top Phrases
############################################################

boeing_top_phrases <- boeing_ngram_counts_final %>%
  slice_max(
    n,
    n = 10
  )

############################################################
# 5H: Create Visualization
############################################################

boeing_ngram_chart <- ggplot(
  boeing_top_phrases,
  aes(
    x = reorder(ngram, n),
    y = n
  )
) +
  geom_col(
    fill = "steelblue"
  ) +
  geom_text(
    aes(label = n),
    hjust = -0.1,
    size = 3
  ) +
  coord_flip() +
  scale_y_continuous(
    expand = expansion(mult = c(0, .10))
  ) +
  labs(
    title = "Top 10 Two-Word Phrases in Boeing Headlines",
    x = "",
    y = ""
  ) +
  theme_minimal()

boeing_ngram_chart

############################################################
# Step 6: Airbus N-Gram Analysis
############################################################

############################################################
# 6A: Choose Phrase Length
############################################################

n_size <- 2

############################################################
# 6B: Load Stop Words
############################################################

data(stop_words)

############################################################
# 6C: Create Airbus N-Grams
############################################################

airbus_ngrams <- airbus_headlines %>%
  unnest_tokens(
    ngram,
    title,
    token = "ngrams",
    n = n_size
  )

############################################################
# 6D: Calculate Stop Word Percentage
############################################################

max_stopword_pct <- 33

airbus_ngrams_filtered <- airbus_ngrams %>%
  mutate(
    words = str_split(
      ngram,
      " "
    )
  ) %>%
  rowwise() %>%
  mutate(
    total_words = length(words),
    stop_words_in_phrase =
      sum(words %in% stop_words$word),
    stop_word_pct =
      stop_words_in_phrase /
      total_words * 100
  ) %>%
  ungroup() %>%
  filter(
    stop_word_pct < max_stopword_pct
  )

############################################################
# 6E: Count Phrases
############################################################

airbus_ngram_counts <- airbus_ngrams_filtered %>%
  count(
    ngram,
    sort = TRUE
  )

head(
  airbus_ngram_counts,
  30
)

############################################################
# 6F: Remove Custom Phrases
############################################################

custom_phrases <- tibble(
  ngram = c(
    "ap top"
  )
)

airbus_ngram_counts_final <- airbus_ngram_counts %>%
  anti_join(
    custom_phrases,
    by = "ngram"
  )

############################################################
# 6G: Examine Final Phrase Counts
############################################################

head(
  airbus_ngram_counts_final,
  30
)

############################################################
# 6H: Select Top Phrases
############################################################

airbus_top_phrases <- airbus_ngram_counts_final %>%
  slice_max(
    n,
    n = 10
  )

############################################################
# 6I: Create Visualization
############################################################

airbus_ngram_chart <- ggplot(
  airbus_top_phrases,
  aes(
    x = reorder(ngram, n),
    y = n
  )
) +
  geom_col(
    fill = "firebrick"
  ) +
  geom_text(
    aes(label = n),
    hjust = -0.1,
    size = 3
  ) +
  coord_flip() +
  scale_y_continuous(
    expand = expansion(mult = c(0, .10))
  ) +
  labs(
    title = "Top 10 Two-Word Phrases in Airbus Headlines",
    x = "",
    y = ""
  ) +
  theme_minimal()

airbus_ngram_chart

############################################################
# Step 7: Display Final Visualizations
############################################################

weekly_plotly

boeing_unigram_chart

airbus_unigram_chart

boeing_ngram_chart

airbus_ngram_chart