Since we are on the topic of searching databases this week, I thought I would share another Google database: Google Trends.
Google Trends is a database of Google web searches and of Google News. Type in a word or words into the search engine and Google generates a chart, similar to Google Ngram, that displays how often these terms have been searched over time on Google. The bottom of the chart shows how often these terms have appeared in Google News.
Since Google Ngram and Corpus Time focuses on books and the written word, Google Trends is a nice complement to searching what ordinary people (i.e. non-authors) have been interested in over the years. Another great aspect of Google Trends is that it ranks regions, cities, and languages in which people have searched for certain terms the most.
I did a Google Trends search on “Barack Obama” and most people searched for “Barack Obama” during the 2008 election. Since his inauguration, interest in Obama has steadied with a peak around the time Bin Laden was killed. After Americans, the Irish search the most for “Barack Obama.” Interestingly, Swedish is the second most language in which people search for Obama.
Google Trends is relatively new and you can only see trends in Google searches since 2004. As time goes on, it will be interesting to see how this tool evolves and what trends in web searches develop over time. Do you think Google Trends is a valid historical methodology? Do people’s web search interests merit historians’ attention?
I feel like in the future this tool could be even more useful to historians then n-gram and “culturomics.” Not only does this also show the frequency in different languages, regions, and cities, it also provides a list of news stories from periods where the search received high frequency so you can see what may have caused that frequency. As a tool like this experiences expanded functionality, or if it is used in tandem with something like the wayback machine (so you can go and mine those news sites to see what else was going on), it has the potential to allow you to explore “culturomics” without many of the pitfalls of n-gram. Through the news stories, you can gain some context, and if they are able to eventually provide a list of top websites visited that could be even more useful. Also showing which regions provided the most searches allows this to be used in a more specific sense then n-gram. Again if they expanded this functionality to allow you to search about specific regions or cities alongside your examination of a word it would be even more useful.
Regarding the question of whether peoples searches merit historians attention, I think they will in the near future. What people search shows what they are interested in, along with what they are hearing about. For the social and cultural historians of the future the data found here could be incredibly useful. The ability to compare the frequency of which people are searching two terms (for example Obama and Romney) in different cities over time has the potential to reveal some interesting things.
I guess what I am saying is that it seems like Google Trends is a neat tool with a lot of potential, and I am glad you showed it to us.
Google Trends seems to fit well into a topic that we’ve been touching on recently in class. As the Web progresses and content is generated specifically for online, historical documents are being created and indexed in real time. Web applications like Google Trends help to make sense of what’s happening at any given time. It’s one of a growing set of analytical tools to help historians index events of historical significance.
The vast amounts of data being created online yields a demand for more robust analytical tools like Google Trends and Ngram. These tools allow historians to dive into large data sets with extremely low time investment as well as enable and encourage further research.
Google Trends — and the Google data associated with it — is a particularly powerful tool for looking forward as well as into the past. Google.org, Google’s philanthropic arm, operates Google Flu Trends based on the Google Trends service. Google Flu Trends predicts flu seasons based on the volume of related searches during times of the year that flu typically manifests.
How can you use Google Trends as a predictive engine for your project?
I want to preface my concerns about Google Trends by saying that I love playing with it, dearly and truly, and I agree with the previous comments about its potential as both a historical and predictive tool. At the same time, my concern about tools like these, while an innovative manipulation of digital data, is that we are constantly forced to measure the good and bad when tools like these are the products of corporations. Google creates tools like these for its own “not evil” agenda of monetizing personal and aggregated usage data and the persistence of those tools depends a great deal on their continuing utility to the company so as to justify the labor involved in maintaining and developing new uses for them. While Google Trends has the potential to be infinitely useful to scholars of all backgrounds as the history of timeline stretches further and further from the present moment, it could also easily be made inaccessible, hidden behind a paywall, or fall into the digital graveyard along with other unloved Google experiments (ahem, Buzz). So while we should obviously seize the opportunity to explore its research applications now, the pessimist in me thinks it would be wise never to count on it being there in the future. It also pays to familiarize yourself with the About section (http://www.google.com/intl/en/trends/about.html) which provides important insight into how the data is actually being manipulated behind the scenes and affirms that you can indeed export the information if you are inclined to tinker with it further on your own.