Google Trends have been around for quite sometime. For non-starters, Google Trends is a public web facility of Google Inc., based on Google Search that shows how often a particular search-term is entered relative to the total search-volume across various regions of the world, and in various languages. One can access Google Trends at the following link-
We thought of analyzing google trends for Analytics in India. Below graph is the result, from 2004 to 2012: –
We had some very interesting finds. The search activity for the term ‘Analytics’ in India was at its peak at around end 2009 – Beg 2010, post which the activity started to cool off. We don’t see any apparent reasons for this. Essentially, the activity around analytics in India is still strong and infact heating up. It would not be apt to correlate the real market activity with search activity.
Gurgaon tops among the major cities in India, which originated search activities on analytics. It was followed by Noida, Bangalore and Hyderabad. This again is contrary to the real market activity. From our research, Bangalore tops the Indian cities with most activity on Analytics, followed by Gurgaon and Mumbai.
Even with relative dip in search activity from India, it remains substantially higher than in US. The average normalized (0-100) search rating from India is 61 vs 32 for US.
We then thought of looking for analytics in India on the newly launched Correlate from Google. Google Correlate (http://www.google.com/trends/correlate) enables you to find queries with a similar pattern to a target data series. In other words: it does what Google Trends does, but in reverse.
The top ten keywords with search activity similar to ‘Analytics’ in India (along with their correlation with “Analytics”) are :-
- 0.9863 google analytics
- 0.9845 ubuntu
- 0.9815 xampp
- 0.9808 centos
- 0.9788 wamp
- 0.9783 video tutorial
- 0.9759 all in
- 0.9749 management studio
- 0.9745 vhd
- 0.9732 ssis
For most of the terms here, the correlation with ‘Analytics’ doesn’t make much of a sense. May be a result of spurious, coincidental correlation in search activity.
Try deep learning using MATLAB