GoogleTrends: Exploring Patterns in Search Data [DBQs & Inquiry]
GoogleTrends allows you to plot the popularity of search terms (since 2004), by geographical region or worldwide. This could be a great way to launch inquiry on a topic in science that has seasonal trends or patterns, and could be used to set up simple DBQ practice. It is limited in the linear presentation of data, and the data are search frequencies rather than scientific data, but as the patterns raise questions, they could be followed-up with searches for more valid sources and explanations.
In the example below, Frank Swain (@SciencePunk) had put in the search term “morning after pill” for the UK and found a peak every Sunday. An interesting pattern that could set up some discussion in class based on reproduction, behaviour, risk management, ethics, hormonal control or more.
This could lead to a quick (though basic) way to set up some simple data-based questions or stimulus for exploration. Here is a plot for the search term “vaccine”. Think of the questions it might raise in discussion.
What questions does it raise and how would it lead to further exploration? Here are some examples:
- Why does it peak each October?
- Why was traffic so high in 2009-10?
- What do you predict for the coming year?
This leads into discussion of sources of information, accessing databases and the reasons for vaccines.
One neat feature is that you can add other search terms to the graph in the same time period, though it will normalise the data. Another is the ‘headlines’ feature that shows some popular news headlines near peaks. Yet another is the ‘predict’ feature that will model the coming year based on trends and patterns. “Predict” is often asked in DBQ’s, so this might make for some good questions. Here’s what happens when we add “flu vaccine”, headlines and the forecast:
From this exploration, you could move onto looking at flu trends, and GoogleTrends has special sections for tracking flu and dengue fever:
This next one is a neat demonstration of what happens when you change the scale of the y-axis. In this case, the second dataset is added, compared to the original and the original becomes much less noticeable as a result. How many times do we tell our students to set appropriate scales on the axes and make use of the space to be able to see trends and patterns?
For another bit of fun, here’s one on “Genome”:
- A quick, easy launching pad for inquiry
- Develop simple DBQs easily
- Does need to be supported by inquiry into more valid sources for the topic
- Each graph is ‘normalised’ which could lead into useful discussion of the effect of scales on data presentation