Google Trends for Marketing Analytics

Required Packages

library(httr2)
library(jsonlite)
library(ggplot2)
library(scales)
library(lubridate)
library(knitr)

Understanding Google Trends Time Options

time_options <- data.frame(
  Time_Code = c("now 1-H", "now 4-H", "today 1-m", "today 3-m",
                "today 12-m", "today 5-y", "all"),
  Meaning = c("Past hour", "Past 4 hours", "Past 30 days",
              "Past 90 days", "Past 12 months", "Past 5 years",
              "Since 2004")
)

kable(time_options)

Time_Code	Meaning
now 1-H	Past hour
now 4-H	Past 4 hours
today 1-m	Past 30 days
today 3-m	Past 90 days
today 12-m	Past 12 months
today 5-y	Past 5 years
all	Since 2004

Tutorial 1: Single Keyword — Brand Search Interest

Function to Pull Google Trends Data

clean_google_json <- function(text) {
  text <- sub("^\\)\\]\\}',?\\n?", "", text)
  fromJSON(text, simplifyVector = FALSE)
}

get_google_trends_direct <- function(keyword, geo, time) {
  
  explore_request <- list(
    comparisonItem = list(
      list(
        keyword = keyword,
        geo = geo,
        time = time
      )
    ),
    category = 0,
    property = ""
  )
  
  explore_request_json <- toJSON(explore_request, auto_unbox = TRUE)
  
  explore_response <- request("https://trends.google.com/trends/api/explore") |>
    req_url_query(
      hl = "en-US",
      tz = "360",
      req = explore_request_json
    ) |>
    req_user_agent("Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36") |>
    req_headers(
      accept = "application/json, text/plain, */*",
      "accept-language" = "en-US,en;q=0.9"
    ) |>
    req_error(is_error = function(response) FALSE) |>
    req_perform()
  
  explore_status <- resp_status(explore_response)
  
  if (explore_status != 200) {
    return(
      list(
        success = FALSE,
        status_code = explore_status,
        data = NULL
      )
    )
  }
  
  explore_text <- resp_body_string(explore_response)
  explore_data <- clean_google_json(explore_text)
  
  widgets <- explore_data$widgets
  
  time_widget <- NULL
  
  for (widget in widgets) {
    if (!is.null(widget$id) && widget$id == "TIMESERIES") {
      time_widget <- widget
    }
  }
  
  if (is.null(time_widget)) {
    return(
      list(
        success = FALSE,
        status_code = NA,
        data = NULL
      )
    )
  }
  
  data_request_json <- toJSON(time_widget$request, auto_unbox = TRUE)
  
  data_response <- request("https://trends.google.com/trends/api/widgetdata/multiline") |>
    req_url_query(
      hl = "en-US",
      tz = "360",
      req = data_request_json,
      token = time_widget$token
    ) |>
    req_user_agent("Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36") |>
    req_headers(
      accept = "application/json, text/plain, */*",
      "accept-language" = "en-US,en;q=0.9"
    ) |>
    req_error(is_error = function(response) FALSE) |>
    req_perform()
  
  data_status <- resp_status(data_response)
  
  if (data_status != 200) {
    return(
      list(
        success = FALSE,
        status_code = data_status,
        data = NULL
      )
    )
  }
  
  data_text <- resp_body_string(data_response)
  data_json <- clean_google_json(data_text)
  
  timeline <- data_json$default$timelineData
  
  dates <- c()
  hits <- c()
  
  for (row in timeline) {
    date_value <- as.Date(
      as.POSIXct(
        as.numeric(row$time),
        origin = "1970-01-01",
        tz = "UTC"
      )
    )
    
    hit_value <- as.numeric(row$value[[1]])
    
    dates <- c(dates, date_value)
    hits <- c(hits, hit_value)
  }
  
  trend_data <- data.frame(
    date = dates,
    hits = hits
  )
  
  list(
    success = TRUE,
    status_code = 200,
    data = trend_data
  )
}

Query Google Trends

trend_result <- get_google_trends_direct(
  keyword = "Starbucks",
  geo = "US",
  time = "today 5-y"
)

if (trend_result$success == TRUE) {
  iot <- trend_result$data
} else {
  iot <- data.frame(
    date = as.Date(character()),
    hits = numeric()
  )
}

Data Status

if (trend_result$success == TRUE) {
  cat("Google Trends data was successfully retrieved.")
} else {
  cat("Google Trends data could not be retrieved from Posit Cloud.")
  cat("\n\nStatus code:", trend_result$status_code)
  cat("\n\nA status code of 429 means Google Trends is rate-limiting requests from this cloud environment.")
}

## Google Trends data could not be retrieved from Posit Cloud.
## 
## Status code: 429
## 
## A status code of 429 means Google Trends is rate-limiting requests from this cloud environment.

Preview the Data

head(iot)

## [1] date hits
## <0 rows> (or 0-length row.names)

Plot Search Interest

if (nrow(iot) > 0) {
  
  ggplot(iot, aes(x = date, y = hits)) +
    geom_line(color = "#00704A", linewidth = 1) +
    geom_smooth(
      method = "loess",
      se = TRUE,
      color = "#1E3932",
      fill = "#D4E9E2",
      alpha = 0.3
    ) +
    scale_x_date(date_labels = "%b %Y", date_breaks = "2 months") +
    labs(
      title = "Google Search Interest: Starbucks",
      subtitle = "United States | Past 5 years | Web Search",
      x = NULL,
      y = "Relative Interest (0-100)",
      caption = "Source: Google Trends"
    ) +
    theme_minimal(base_size = 12) +
    theme(
      plot.title = element_text(face = "bold", size = 14),
      plot.subtitle = element_text(color = "gray50"),
      axis.text.x = element_text(angle = 30, hjust = 1)
    )
  
} else {
  
  plot.new()
  text(
    x = 0.5,
    y = 0.5,
    labels = "Google Trends request was rate-limited by Posit Cloud.\nStatus code: 429",
    cex = 1.1
  )
  
}

References

Google Trends. (2024). Google Trends Tutorials [YouTube series]. Google Search Central. https://developers.google.com/search/blog/2024/09/google-trends-tutorials

Massicotte, P., & Eddelbuettel, D. (2023). gtrendsR: Perform and Display Google Trends Queries [R package]. https://cran.r-project.org/package=gtrendsR

Chan, M. (2019). Vignette: Google Trends with the gtrendsR package. https://martinctc.github.io/blog/vignette-google-trends-with-gtrendsr/