Christian Andrews
December 8, 2017
It was the best of times, it was the worst of times… depending on which side you were on. The best for Google, the worst for pretty much everyone else.
After the great Search Engine Wars of the late 20th and early 21st century, Google.com emerged victorious. Now serving well over 2 Trillion searches per year. That is roughly 267 queries per year for every man woman and chile on Earth. Almost 1 per day per person.
Google is officially a verb. https://www.merriam-webster.com/dictionary/google
Suffice it to say, Google is a major play in our digital world and our modern society.
If you want to know what is happening in the world, look no further than Google Trends. https://trends.google.com/trends/
Google has opened up access to the Key Words used in searches made by those around the world. Let’s take a look at a fairly recent trend: “Internet of Things” since 2010
phrase="internet of things"
trend_data<-gtrends(phrase,"US",time = "2010-01-01 2017-12-08")$interest_over_time
plot(trend_data$hits, type = 'l',ylab='Hits per day',xlab='Date Range',xaxt='n')
Trends are great but, what is missing? The forecasting!
Below you will see the same search trend but with a Holt Winters forecast. Now you can see, if you should quit your day job and get into the IoT revolution or not.
tdata<-ts(trend_data$hits, frequency = 24)
tforecasts <- forecast(HoltWinters(tdata), h=22,level=0.55)
plot(tforecasts,ylab='Hits per day',xlab='Date Range',xaxt='n')
Here is a closer look at a related query for “iot” over the past 3 months. Judging from the highly cyclic, 7-day trend, the term is much more popular during the week, than on the weekends. So don’t strike up an “iot” conversation at the bar on Saturday night…
phrase="iot"
trend_data<-gtrends(phrase,"US",time = "today 3-m")$interest_over_time
tdata<-ts(trend_data$hits, frequency = 7)
tforecasts <- forecast(HoltWinters(tdata), h=22,level=0.55)
plot(tforecasts,ylab='Hits per day',xlab='Date Range',xaxt='n')