Wednesday, April 14, 2010

Google Search Trends for "Bahai"

The recent Baha'is: Forecasts I article here featured data available on the internet in its methodology and the need to study trends as a basis for successful forecasting. What might be learned from Google's marvelous search trend service? Interesting and somewhat unexpected findings are presented and discussed.

Starting with the single search word "baha'i", Google reports insufficient data to apply its analysis technology. However, the word "bahai" yielded results.

Time Trends 2004 to Present
Due to some technical limitations in this shop and the narrow width of text on this web page, the time trend is presented in the following two images, which are parts of one graph horizontally:
The "Search Volume Index" shown above is scaled where 1.00 is an average of an initial period in the data series. We see that over 2004 to 2006, the search volume for Baha'i was relatively constant or perhaps slowly drifting downward, although the author has not done the number crunching to evaluate the statistical significance, if any, for this downward sloping trend in this period. [Please contact the author by private email if you would like a 15K .csv file of all the Google data presented in this report.]

The lower graph in the data series, News reference volume, shows various spikes which are most often related to Baha'i news coming out of Iran. Much of this activity probably originates with press releases by Baha'i institutions.
The above continuation of the same graph shows a more pronounced drop in "bahai" search volume. Indeed, by 2010, searches decline to less than one half of the initial period starting in 2004. The dramatic down-trend might be of considerable concern to many Baha'is, who might prefer to see an up-trend result. Of course, correct interpretation of this data is a difficult task; but on the face of it, the down-trend suggests decreased interest in the "bahai" keyword.

One hypothesis might be that the Volume trend may substantially be a function of the rate at which new Baha'i content is added to the internet. We might reason that if such content is rising rapidly, most searchers, whether Baha'is or others including "seekers", would be more likely to use search engines to discover new content.

The points above labelled A through F by Google correspond to news stories, with A denoting Baha'i-related news in Egypt and B to F, in Iran, respectively. Generally, labelled times A-F on the Search Volume plot correspond to increases or even spikes in the News plot. This may help evaluate the meaning of Search Volume data -- to some extent it relates to the amount of Baha'i-related News in the secular media.

The above data consisting of two variables over time, Volume and News, may be used to construct composite variables defined as some derivative of the two raw input variables. One such variable might be constructed based on Volume minus News to explore the information further, regarding volume increases not associated with secular news, and hence, perhaps indicative of interest in internal events in the Baha'i community not reported in the secular media.

Mysterious 2007 Search Volume Spike
Without mathematically constructing and plotting such a composite variable, what can we note via visual inspection? First, many spikes in Volume seem to be associated with some rise in the News variable. This might be seen as a normal response among Baha'is to search the internet for further information related to a Baha'i news story in the secular media.

Now let us focus on the biggest Volume spike, near the end of the first half of 2007. It might seem strange that the biggest short-term increase in "bahai" search volume over some seven years (2004-2010) is not associated with any noteworthy rise in the News variable. In other words, this search Volume spike may suggest that something happened in the Baha'i community which was of great interest and not reported in the secular news media. What happened?

Could it be the Momen Moment, described by this reporter? Or something else causing an apparent stampede, most probably by Baha'is, to search the internet ... regarding what? More on this theme below.

Baha'i Search by Country
It may be a surprise that the United States, with all of its internet prowess, beats out only Chile and India in this top ten list. Over the last seven years, most "bahai" search on Google comes out of Iran and Israel. To this reporter, this result emphasizes how much people, likely mostly Baha'is, in Iran (and we might include Malaysia, too) use the internet to learn of Baha'i-related matters.

The number one spot of Iran in this data might prompt us to reflect on the interpretation of the Volume data above. Namely, if we suppose that some censorship, perhaps starting in the 2007-2008 period, limited the ability of Google searchers in Iran to look up the "bahai" keyword, it might explain some or even all of the Volume drop seen in the first graphs up to the present. This consideration highlights how wrong data interpretations can be. For example, if this possible censorship factor exists as described and is removed from the data, the Volume trend might appear to be flat or even rising, and lead to very different conclusions.

This country data also brings to mind that Baha'i institutions might strive to make their web pages simple; that is, considering their world-wide audience, and the likelihood that many site visitors may have slow computers or slow internet connections, or both, if web page byte count from image and/or html size is kept as low as possible, the pages will not take forever to load and be viewed.

This is the first venture by the author in Google trend analysis. It may be possible to generate the above graphs excluding certain countries, such as Iran and Israel (mostly Haifa as shown below), to see if this Google tool might be able to screen the data in a way that might reveal trending interest in Baha'i matters by various sub-populations.

In this regard, please note also that only one keyword "bahai" was used. The Google tools are capable of much more detailed trend analysis, including multiple keywords, assuming sufficient sample size is found to document the trend with sufficient confidence in result accuracy.

Baha'i Search by City
Haifa, Israel, the location of the Baha'i World Centre, is far and away the top city on earth conducting internet search for "bahai" according to Google, over the seven year data series considered. When we add Petah Tiqwa, Israel, the above graphic shows that about one half of all "bahai" internet searches originate in Israel.

Clearly, Baha'i munchkins in Haifa have been very busy on the internet. And they can use this data to show supervisors that they are not just playing computer games all day.

Since the Baha'i World Centre is itself a center of Baha'i information, such as computerized files of Baha'i writings, world-wide administrative information and such, one might wonder what they are searching for. To say the obvious, one searches for information one does not already have. For example, news of activities in Baha'i communities might be of great interest as local press releases become new internet content.

On the other hand, many Baha'is (excluding the author), whatever their membership, and others might see this data as an occasion for a paranoid fit -- make a peep on the internet and your every syllable may be quickly logged at the Baha'i World Centre. Talk about information-gathering power. Can any of the various Baha'i "splinter groups" compare with this? Not a chance, according to the present data at least. Now we know one more factor regarding who the winning horse currently is and no doubt will be going forward.

All else being equal, this search breakdown by city suggests that approximately one half of the big Volume spike in mid-2007, which this report has suggested might reflect an orchestrated publicity event among Baha'is, dubbed as the Momen Moment, most probably came from Baha'is in Israel. Was this just routine monitoring of the action, so to speak? Might it also be consistent with the supposition that some possible liberal conspirators, if they exist, were located at the Baha'i World Centre? Alas, this data will not tell us that, so this remains only as an interesting clue.

On the other hand, the other half of the supposed Momen Moment volume spike in searches was distributed widely around the world. In sum, the observed 2007 Volume spike, if in fact related to the Momen Moment, may partially quantify success of the imagined liberal conspirators described in that report.

As for the remainder of the entire world beyond the Baha'i search traffic in Israel, almost all searchers are located in major cities in Canada (3 cities),Australia (3 cities) and the USA (2 cities).

In conclusion, the analytical tools and the images above graciously provided by Google might be used to further research behavior of Baha'is and perhaps the general public also, with proper data screening, regarding the ever-growing presence of Baha'i material from all corners on the internet.
© 2010 James J Keene

2 comments:

  1. Thanks James. The Google trends might be misleading. Terms like "spirituality", "religion", "christianity" and "human rights" also trend downward for the same period. Perhaps given the explosive growth of the Internet over the past few years the volume of searches for such terms are being overwhelmed by the growth in total number of searches (ie, there are a lot of other things that people are searching for now as opposed to a few years ago). Noticed that "happiness" did trend upward though. Fascinating what this says about the human condition.

    ReplyDelete
  2. Re "The Google trends might be misleading," agreed, at least without careful analysis to quantify independent factors in the data.

    Re "spirituality", "religion", "christianity" and "human rights" also trend downward for the same period, this welcome observation might suggest that factor analysis, partial correlation, etc, might remove from Bahai-time covariance (the downward slope) some general effects you identified.

    Re your "explosive growth of the Internet" comments, as I understand it, the Google analysis looks at "within search-word group" variance in the Bahai-time plot, which would not depend on the size of other search-word groups or overall number of internet users. It may be correct, then, that regardless of the number of potential searchers, search for "bahai" has decreased, the opposite of what increased number of internet users might predict. That is, on a Search-Per-Internet-User basis, the downward slope would seem to be even more dramatic.

    But your point may be important concerning a change in the composition of searchers, a sort of sample incontiguity. That is, the new internet users may be quite different than previous users which is really a mess, because we have a sample of, say, Group A to the left in the time plot, and Group B to the right.

    If reports surface of further trend analysis, we might have a fun time, each new report looking at a different angle, attempting to control for confounding variables, etc. Do we have a Master's Thesis waiting to be written here?

    With "Noticed that 'happiness' did trend upward", thank Anonymous for an interesting hint for marketing the Baha'i message. Hey, maybe I should add a description of this blog like "find happiness here" or something.

    ReplyDelete