Not exceeding a year, Google Trends data was made accessible in real time; and progressively, it’s helping people around the world explore the global reaction to major events.
The massive volume of searches — trillions take place every year — make Google Trends one of the world’s largest real time datasets. Scrutinizing what individuals quest for delivers a unique perspective on what they are currently interested in and curious about.
Consequently as soon as a huge news story happens, how can you best interpret this data?
What is Trends data?
Data from Google Trends is a balanced sample of our Google search data. It’s anonymized (no one is personally identified), categorized (determining the topic for a search query) and aggregated (grouped together). As a result it allows us quantify interest in a particular topic across search, from around the globe, right down to city-level geography.
The Google Trends free data explorer enables you search for a particular topic on Google or a specific set of search terms. Make use of the implement and you can see search interest in a topic or search term over time, where it’s most-searched, or what else people search for in connection with it.
There are two ways to filter the Trends data: real time and non-real time. Actual time is an arbitrary sample of searches from the last seven days, while non-real time is another random sample of the full Google dataset that can go back anywhere from 2004 to ~36 hours ago. The graphs will display you either one or the other, but not both together, because these are two separate random samples. We take a sample of the trillions of Google searches, because it would otherwise be too large to process quickly. Via selecting our data, we can look at a dataset representative of all Google searches, while finding insights that can be processed within minutes of an event happening in the real world.
It’s a unique and powerful dataset, which can complement others, like demographic data from the census, as shown here in the Washington Post. By means of sampling, it gives us a way to analyse what people are searching for in real time as events unfold. But combining data can be tricky — for instance, it doesn’t make sense to compare Trends to other Google datasets, which are measured in different ways. For example, AdWords is meant for insights into monthly and average search volumes, specifically for advertisers, while Google Trends is designed to dig further into more granular data in real time.
What do the numbers mean?
Trends is a potent tool for storytelling because it can allow us to explore the magnitude of different moments and how people react to those moments. We can look back and compare different terms against each other, like how different sports have ranked since 2004. Furthermore we have the ability to take the total searches for an event to help understand its sheer magnitude. The Moment we released our 2015 Year in Search, we found there were astoundingly over 439 million searches on Google when Adele came back with ‘Hello’.
What’s most useful for storytelling is our normalized Trends data. It simply demonstrates that anytime we look at search interest over time for a topic, we’re looking at that interest as a proportion of all searches on all topics on Google at that time and location. Whenever we glance at regional search interest for a topic, we’re looking at the search interest for that topic in a given region as a proportion of all searches in pertinence to every topic on Google in that same place and time.
You May Also Like: How To Recover Deleted Files From Mac
Instantaneously, when we look at the Trends around Bernie Sanders, we can see that Vermont has the highest search interest in the current senator. This is because of all states, Vermont has the highest percentage of searches for Sanders out of all searches in that state. If we had looked at raw data rather than normalized values, we would’ve seen larger states with higher populations rise to the top of the ranks.
The standardization is extremely important: the number of people searching on Google changes constantly — in 2004 search volume was much smaller than it is today, so raw search numbers wouldn’t give you any way to compare searches then and now. By normalizing our data, we can make deeper insights: comparing different dates, different countries or different cities.
The context of our numbers also matters. Our data is Indexed to 100, where 100 is the maximum search interest for the time and location selected. This simply means that that if we look at search interest in the 2016 elections since the start of 2012, we’ll see that March 2016 had the highest search interest, with a value of 100.
If we look at search interest in only March 2016, though, we can see that March 16 has the highest search interest, because we’ve re-indexed our values for just that month.
How do you put the numbers in context?
Since Google Trends data is presented as an index, we often get the question: “how important is this?”
There are a few ways to assess this. The first step is to understand relative search interest in the topic compared to itself — or what we would call a “spike”.