Taylor Swift and Travis Kelce’s connection exemplifies how two disparate areas of popular culture can intersect and be widely covered in the media. Known as a famous musician, Swift belongs to one of the most important stars in the music industry while Kelce has earned recognition in professional sports. As their affair became known in 2023, it formed an important linkage between entertainment and professional sports fields. Their relationship has attracted attention from readers who may normally not pay attention to news from both areas and has generated numerous opportunities for sports and entertainment media to reach new audiences.
This project aims to assess the way media coverage related to Taylor Swift and Travis Kelce has evolved from September 2023 to September 2024. Instead of analyzing the relationship between the two individuals, the objective will be to understand how often the news agencies published articles on them and whether the fact of their relationship significantly increased the number of articles published. By analyzing the number of articles published, this project will give a basic understanding of the relationship between celebrity culture, sports media, and entertainment journalism.
The main question that we will be addressing in this research is: How has the media coverage on Taylor Swift and Travis Kelce evolved between September 2023 and September 2024? This is an opportunity to study how an event involving two persons of different professions creates a common topic for media discussion and how it attracts the attention of various news organizations.
The source of data in this work is Media Cloud, an online medium that enables individuals to review coverage of news and determine how often a certain topic is mentioned in published articles. Media Cloud provides an efficient tool for analyzing the development of the media focus of attention since the searches can be organized according to keywords and can be assessed over a particular time frame. In relation to this project, the website will be used to find the number of articles associated with Taylor Swift and Travis Kelce.
The data is organized with the help of two search queries. The first search query focuses on looking for articles that mention both Taylor Swift and Travis Kelce. This is done in order to find the articles which mention both personalities in the text. The second search focuses on the articles mentioning Travis Kelce in a broader sense without a reference to Taylor Swift. Making a comparison of these two searches allows for measuring the amount of coverage associated with Taylor Swift.
The stated period extends from September 1, 2023 to September 1, 2024. This time frame encompasses the initial 12 months of public scrutiny of their romance and entails key events that must have affected how the media covered them, including Swift going to the Chiefs’ games, the surge of attention to Kelce’s career in the NFL, and the ongoing discussions of their romance in the media, both in entertainment and sports.
The API provides the information about how many stories appeared on each date. Then these numbers are grouped into weekly totals so that changes in the media attention are easier to distinguish. The weekly figures have the added bonus of allowing comparisons between different time periods without paying attention to certain days with high or low media attention rate.
api_key <- Sys.getenv("MEDIACLOUD_KEY")
if (api_key == "") {
stop(
"API key not found. Add MEDIACLOUD_API_KEY to your .Renviron file."
)
}
base_url <- "https://search.mediacloud.org/api/search/count-over-time"
start_date <- "2023-09-01"
end_date <- "2024-09-01"
platform_name <- "onlinenews-mediacloud"
collection_id <- "34412234"
The following function sends a search request to Media Cloud and retrieves the number of stories matching each search query.
get_media_counts <- function(search_query, variable_name) {
response <- request(base_url) |>
req_headers(
Authorization = paste("Token", api_key),
Accept = "application/json"
) |>
req_url_query(
platform = platform_name,
q = search_query,
start = start_date,
end = end_date,
cs = collection_id
) |>
req_timeout(60) |>
req_perform()
response_data <- resp_body_json(
response,
simplifyVector = TRUE
)
counts <- response_data$count_over_time$counts
if (is.null(counts) || length(counts) == 0) {
stop("No count data returned. Check the API response and search query.")
}
as_tibble(counts) |>
transmute(
date = as.Date(substr(date, 1, 10)),
stories = as.integer(count)
) |>
rename(!!variable_name := stories)
}
We can now retrieve the two sets of media coverage data.
both_query <- '"Taylor Swift" AND "Travis Kelce"'
kelce_query <- '"Travis Kelce"'
both_data <- get_media_counts(
both_query,
"both_mentions"
)
kelce_data <- get_media_counts(
kelce_query,
"kelce_mentions_total"
)
head(both_data)
## # A tibble: 6 × 2
## date both_mentions
## <date> <int>
## 1 2023-09-01 0
## 2 2023-09-02 0
## 3 2023-09-03 0
## 4 2023-09-04 0
## 5 2023-09-05 0
## 6 2023-09-06 1
The two datasets are combined using their dates. Missing counts are replaced with zero.
A third variable is created to estimate stories mentioning Travis Kelce without also mentioning Taylor Swift. This is calculated by subtracting the number of stories mentioning both individuals from the total number of stories mentioning Kelce.
daily_data <- full_join(
both_data,
kelce_data,
by = "date"
) |>
mutate(
both_mentions = replace_na(both_mentions, 0L),
kelce_mentions_total = replace_na(kelce_mentions_total, 0L),
kelce_only_mentions = pmax(
kelce_mentions_total - both_mentions,
0L
)
) |>
arrange(date)
weekly_data <- daily_data |>
mutate(
week_start = floor_date(
date,
unit = "week",
week_start = 1
)
) |>
group_by(week_start) |>
summarise(
both_mentions = sum(both_mentions),
kelce_mentions_total = sum(kelce_mentions_total),
kelce_only_mentions = sum(kelce_only_mentions),
.groups = "drop"
) |>
arrange(week_start)
head(weekly_data)
## # A tibble: 6 × 4
## week_start both_mentions kelce_mentions_total kelce_only_mentions
## <date> <int> <int> <int>
## 1 2023-08-28 0 65 65
## 2 2023-09-04 2 373 371
## 3 2023-09-11 74 285 211
## 4 2023-09-18 205 304 99
## 5 2023-09-25 786 929 143
## 6 2023-10-02 548 664 116
The following table displays the first several observations in the dataset. Each row represents one week of media coverage.
kable(
head(weekly_data, 10),
caption = "First Ten Weeks of Media Coverage"
)
| week_start | both_mentions | kelce_mentions_total | kelce_only_mentions |
|---|---|---|---|
| 2023-08-28 | 0 | 65 | 65 |
| 2023-09-04 | 2 | 373 | 371 |
| 2023-09-11 | 74 | 285 | 211 |
| 2023-09-18 | 205 | 304 | 99 |
| 2023-09-25 | 786 | 929 | 143 |
| 2023-10-02 | 548 | 664 | 116 |
| 2023-10-09 | 377 | 579 | 202 |
| 2023-10-16 | 311 | 450 | 139 |
| 2023-10-23 | 324 | 458 | 134 |
| 2023-10-30 | 195 | 342 | 147 |
The dataset includes the starting date of each week, the number of stories mentioning both individuals, the total number of stories mentioning Travis Kelce and the estimated number of stories mentioning Kelce without Taylor Swift.
The next table provides basic descriptive statistics for the collected data.
summary_table <- weekly_data |>
summarise(
total_weeks = n(),
total_joint_mentions = sum(both_mentions),
total_kelce_mentions = sum(kelce_mentions_total),
average_joint_mentions = mean(both_mentions),
highest_joint_mentions = max(both_mentions),
average_kelce_mentions = mean(kelce_mentions_total)
)
kable(
summary_table,
digits = 2,
caption = "Summary of Media Coverage"
)
| total_weeks | total_joint_mentions | total_kelce_mentions | average_joint_mentions | highest_joint_mentions | average_kelce_mentions |
|---|---|---|---|---|---|
| 53 | 11789 | 18938 | 222.43 | 1272 | 357.32 |
These statistics provide an overview of the amount of media attention recorded during the study period. The totals represent the number of matching stories counted by Media Cloud, while the averages show the typical weekly volume.
The following graph compares weekly stories mentioning both Taylor Swift and Travis Kelce with the broader number of stories mentioning Travis Kelce.
weekly_long <- weekly_data |>
select(
week_start,
both_mentions,
kelce_mentions_total
) |>
pivot_longer(
cols = -week_start,
names_to = "coverage_type",
values_to = "story_count"
) |>
mutate(
coverage_type = recode(
coverage_type,
both_mentions = "Taylor Swift and Travis Kelce",
kelce_mentions_total = "Travis Kelce overall"
)
)
ggplot(
weekly_long,
aes(
x = week_start,
y = story_count,
color = coverage_type
)
) +
geom_line(linewidth = 1) +
labs(
title = "Weekly Media Coverage of Taylor Swift and Travis Kelce",
subtitle = "September 2023 through September 2024",
x = "Week",
y = "Number of News Stories",
color = "Search Category"
) +
scale_y_continuous(labels = comma) +
theme_minimal() +
theme(
plot.title = element_text(face = "bold", size = 15),
legend.position = "bottom"
)
This graph shows how the volume of media coverage changed throughout the study period. Increases and decreases in the lines can help identify periods when news organizations published more stories about the individuals.
The following table identifies the weeks with the highest number of stories mentioning both Taylor Swift and Travis Kelce.
highest_coverage <- weekly_data |>
arrange(desc(both_mentions)) |>
slice_head(n = 10)
kable(
highest_coverage,
caption = "Ten Weeks With the Highest Joint Media Coverage"
)
| week_start | both_mentions | kelce_mentions_total | kelce_only_mentions |
|---|---|---|---|
| 2024-02-05 | 1272 | 1845 | 573 |
| 2024-02-12 | 882 | 1288 | 406 |
| 2024-01-29 | 862 | 1182 | 320 |
| 2023-09-25 | 786 | 929 | 143 |
| 2023-10-02 | 548 | 664 | 116 |
| 2023-10-09 | 377 | 579 | 202 |
| 2024-01-22 | 374 | 627 | 253 |
| 2023-10-23 | 324 | 458 | 134 |
| 2023-10-16 | 311 | 450 | 139 |
| 2023-11-13 | 294 | 475 | 181 |
These weeks represent the periods with the greatest volume of stories matching the joint search query. They can help identify important moments for further investigation.
The results indicate that there have been significant shifts in the media focus on Taylor Swift and Travis Kelce from September 2023 through September 2024. The most striking pattern observed was the rise in media coverage during late September and early October 2023. The coverage rose again at the end of January, with the week of February 5, 2024 being the week with the highest amount of media coverage of Taylor Swift and Travis Kelce together. During that week, there were 1,272 stories talking about both Taylor Swift and Travis Kelce, while 1,845 stories featuring only Travis Kelce were published.
After the peak observed in February, the rate of media coverage dropped significantly and remained very low for a long period of time. There were some minor instances of media coverage rising in spring and summer, but these rises were not as significant as the one that happened in February. The graphs demonstrate that the trends registered in the two categories were very similar because when media coverage of Travis Kelce rose, the media coverage mentioning both of those figures increased too.
The cleaned weekly dataset is exported as a CSV file so it can be used in future analysis.
write_csv(
weekly_data,
"travis_kelce_taylor_swift_media_coverage.csv"
)
cat("Dataset successfully exported.")
## Dataset successfully exported.
The dynamic between Travis Kelce and Taylor Swift creates a unique chance to explore the intersection of sports, entertainment, and celebrity culture. Both Swift and Kelce had huge public presences before their relationship blossomed, but now they are connected in a way that grabs the interest of various sectors of media.
The first stage of this research project sets the stage for analyzing media attention using Media Cloud. The team will analyze weekly media coverage and identify times with significant spikes in coverage to see how the couple’s media attention evolved during the first year of their relationship. Going forward, the project will utilize data to understand the unique intersection of celebrity culture and sports and the role that these famous people can play in shaping the coverage in different fields.