Online news outlets published more coverage of Republican presidential candidate Donald Trump than of Democratic rival Kamala Harris throughout most of the 2024 election, an analysis of GDELT data shows.

This chart compares daily counts of online articles mentioning “Donald Trump” and “Kamala Harris” between July 1, 2024, shortly before Democratic Harris entered the race, and Nov. 6, 2024, the day Harris conceded the race to Trump. “Volume” values are the number of articles published per day in all U.S.-based online news outlets tracked by GDELT.

The agenda setting theory of mass communication suggests that media coverage volume about topic can influence the amount of attention the public pays to the topic. Other factors may affect the development of public opinions about the topic, though.

Code

Below is the R script that gathered the data and produced the chart. If you’re feeling brave, try fiddling with the queries.

# ------------------------------
# Install and load required packages
# ------------------------------

if (!require("tidyverse"))
  install.packages("tidyverse")
if (!require("plotly"))
  install.packages("plotly")
if (!require("readr"))
  install.packages("readr")

library(tidyverse)
library(plotly)
library(readr)


# ------------------------------
# Define date range
# ------------------------------

startdate <- "20240701"
enddate   <- "20241106"


# ------------------------------
# Fetch Topic A: Donald Trump
# ------------------------------

query <- "'Donald Trump' SourceCountry:US"

vp1 <- "https://api.gdeltproject.org/api/v2/doc/doc?query="
vp2 <- "&mode=timelinevolraw&startdatetime="
vp3 <- "000000&enddatetime="
vp4 <- "000000&format=CSV"

text_v_url <- paste0(vp1, query, vp2, startdate, vp3, enddate, vp4)
v_url <- URLencode(text_v_url)

VolumeA <- tryCatch({
  # Try API request
  Volume <- read_csv(v_url)
  Volume$Date <- as.Date(Volume$Date, "%Y-%m-%d")
  Volume <- Volume %>% filter(Series == "Article Count")
  Volume
}, error = function(e) {
  message("⚠️ Trump query failed — loading from GitHub backup")
  # Download from GitHub and process
  Volume <- read_csv("https://github.com/drkblake/Data/raw/refs/heads/main/TrumpResultsGDELT.csv")
  Volume$Date <- as.Date(Volume$Date, "%Y-%m-%d")
  Volume <- Volume %>% filter(Series == "Article Count")
  Volume
})


# ------------------------------
# Fetch Topic B: Kamala Harris
# ------------------------------

query <- "'Kamala Harris' SourceCountry:US"

text_v_url <- paste0(vp1, query, vp2, startdate, vp3, enddate, vp4)
v_url <- URLencode(text_v_url)

VolumeB <- tryCatch({
  # Try API request
  Volume <- read_csv(v_url)
  Volume$Date <- as.Date(Volume$Date, "%Y-%m-%d")
  Volume <- Volume %>% filter(Series == "Article Count")
  Volume
}, error = function(e) {
  message("⚠️ Harris query failed — loading from GitHub backup")
  # Download from GitHub and process
  Volume <- read_csv("https://github.com/drkblake/Data/raw/refs/heads/main/HarrisResultsGDELT.csv")
  Volume$Date <- as.Date(Volume$Date, "%Y-%m-%d")
  Volume <- Volume %>% filter(Series == "Article Count")
  Volume
})


# ------------------------------
# Merge datasets
# ------------------------------

VolumeAB <- merge(VolumeA, VolumeB, by = "Date")
VolumeAB$VolumeA <- VolumeAB$Value.x   # Trump
VolumeAB$VolumeB <- VolumeAB$Value.y   # Harris


# ------------------------------
# Create interactive plot
# ------------------------------

fig <- plot_ly(
  data = VolumeAB,
  x = ~Date,
  y = ~VolumeA,
  name = 'Trump',
  type = 'scatter',
  mode = 'lines',
  line = list(color = "#AE2012")
) %>%
  add_trace(
    y = ~VolumeB,
    name = 'Harris',
    mode = 'lines',
    line = list(color = "#005F73")
  ) %>%
  layout(
    title = 'U.S. coverage volume',
    xaxis = list(title = "Date", showgrid = FALSE),
    yaxis = list(title = "Volume", showgrid = TRUE)
  )

fig


# ------------------------------
# Save merged dataset
# ------------------------------

write_csv(VolumeAB, "VolumeAB.csv")

Fiddling with the queries

Feeling brave? Try finding the code line that reads:

query <- "'Donald Trump' SourceCountry:US"

… and replacing it with:

query <- "('Donald Trump' OR 'Trump administration') AND 'election' SourceCountry:US"

… and finding the line that reads:

query <- "'Kamala Harris' SourceCountry:US"

… and replacing it with:

query <- "('Kamala Harris' OR 'Harris campaign') AND 'election' SourceCountry:US"

The resulting chart looks different, doesn’t it? Why? Here’s a hint: The revised queries use “OR” and “AND” as boolean logic terms that specify combinations of criteria that must be true in order for a given article to be selected for inclusion in the chart.

You might even try other sets of query terms about Trump and Harris. In fact, you can get away from Trump and Harris altogether and compare other queries for other time periods. Here are the code lines you would need to tweak:

startdate <- "20240701" # To change the starting date
enddate   <- "20241106" # To change the ending date

# AND / OR:

name = 'Trump', # To rename the red line
name = 'Harris', # To rename the blue line

# AND PROBABLY:

title = 'U.S. coverage volume', # To change the chart's title