Tuesday, January 2, 2018

The vox populi of Google Trends

In What are people interested in? I explained the background and methodology I have used to exploit Google Trends to understand what people are interest in based on what they are searching about. The drawbacks and advantages of this approach are described in that post.

These are the results for the 97 terms which I searched.

Click to enlarge.

More legibly, here are the results broken down into arbitrarily chosen ranges.

Red - Index score of 200 - 500. Search items that are frequently searched.
Yellow - Index score of 50 - 100.
Dark Green - Index score of 25 - 50.
Green - Index score of 10 - 25.
Light Green - Index score of 5 - 10.
Pale Green - Index score of 1 - 5.
Light Blue - Index score of 0.1 - 1
Dark Blue - Index score below 0.1

Click to enlarge.

Click to enlarge.

Click to enlarge.

Click to enlarge.

Click to enlarge.

Click to enlarge.

Click to enlarge.

Click to enlarge.

The Column Headings
Index Average (2004 - 2017) - The average of the monthly index for the entire 14 year time frame. Gives an absolute sense of relevance/importance compared to the other 96 search items. For example, people search on "home" about 2.5 times as often as they do "family".

SD - The calculated standard deviation. The amount of variance in the monthly index over the fourteen year period. Very low SD numbers indicate little variance in search interest over the time frame. Very high SD numbers indicate large surges and retreats in interest.

SD/AVG - The SD number as a percentage of the base index average for that search term. Normalizes the reference.
Even though "Home" has relatively high SD number of 64.2, this is against the highest base of 490.2. SD/AVG is actually 0.1 (10%) around the average which is not a lot of variance.

Trendline (Last 12 mo to First 12 mo) - The average index number for the last twelve months (2017) over the first 12 months (2004). This provides an indication of relative rise or fall in interest over time. Interest in "Home" is 1.0, that is to say that it is as strong in 2017 as it was in 2004. However, interest in "War" has fallen to 60% of what it was in 2004. Not unexpected given that 2004 marked the initial phases of the Second Gulf War (2003 onwards).

Trendline (half on half) - The average index for the search item for the second half of the 2004-2017 time period compared to the first half. Since the time periods are longer, this smooths out any noise from just taking a 12 month view (beginning and end). Should reflect more systemic changes in interest in the topic.

Percent of Total Attention - The total of all indexes of interest over the 97 search items is 2,784. Percent of Total Attention shows the relative share of searches of this item against all searches. For example, 17.6% percent of all searches were on the item "Home."

Ratio as a % of Home - The total of this particular search item versus the most searched item, "Home." Provides another sense of relative importance. For example, people, based on their searches, are only 20.4% as interested in "Education" as they are in "Home."

CAVEATS, CAVEATS, CAVEATS!

See What are people interested in? In addition, it is worth adding, that just because people are searching on something, or not, is no indication of moral or absolute value judgement. For example, just because the search index for "Global Warming" is 6.4 (compared to 490.1 for "Home") and they are searching on it less frequently (down 60%) than they used to do, does not mean global warming isn't real, does not mean that it isn't important to people, and it does not mean that it isn't consequential. What it tells us is more limited. It tells us that people are not searching the term nearly as much as they used to do and that global warming occupies a quite small percentage of people's attention (as reflected by searches).

We can make up innumerable stories as to why that might be and those stories might or might not be plausible. The numbers give us visibility into the attention economy, they provide no direct hint as to why the numbers are what they are.

No comments:

Post a Comment