While looking for new ways to play with Word Clouds, I did the cloud from the previous post on the SEOMoz blog main page. I have been using the clouds with keyword research as well, filtering on types of words, number of searches, etc.
I wanted to find some public (not proprietary client data) to play with, so I thought about News. Some of my hobbies revolve around researching my Ancestry and Scrapbooking. I love the idea of preserving information for future generations.
So, for the last couple of days, I have been pulling the keywords from the NYT Homepage. I exclude words that are related to date, brand name and function of the page and concentrate on keywords that are repeated 4 or more times (less than that and the cloud is very confusing!).
Following are word clouds for the last 3 days.
What do these images “tell us?” I’m not really sure. Right now, they’re just a snapshot in words of what was being reported/talked about on those days.
When Tableau 8 releases, and word clouds like this can be uploaded to Tableau Public – where interaction with the data is possible, these might be more fun. Like looking at the frequency of certain words over time, looking for trends in weeks or months of data. It could be the basis for some interesting sociological research. Or it could just be fun, like looking at old snapshots can be.
I’ve really gotten interested in Word Clouds since I’ve been playing around with the beta version of Tableau 8. It’s just so fun to see the words that POP in both size and color…
The following image is a keyword cloud using words from the SEOMoz blog homepage today, 25 Feb 2013. It’s been filtered to exclude date related terms and terms related to the functioning of the blog (like post, read full, comment, etc.). The size of the words is the average number of repeats on the site. The color shows whether the keywords are just in the content or if they appear in the page title, meta-description (or both) on the site. The image has also been filtered to show words with 3 or more repetitions (otherwise the volume of keywords makes it harder to see any patterns.)
Using a word cloud can help you visually focus on the words that are being used most often on a site. And compare that to the words you WANT the search engines to see on the site.
There are lots of variations of this keyword/word cloud theme… you could look at the Google AdWords keyword tool results of a scan of your site. Or, the keywords in your analytics account that show the keywords that are actually generating traffic to your site. Or the advertising keyword results via Google AdWords or Bing AdCenter.
Ah, the possibilities are endless! 🙂
I love data. I really love turning data into information, going from columns of words and numbers into something that you can actually make sense of …
For example, here are the keywords that brought organic search traffic to my website in 2012 (excluding, of course the ubiquitous not provided and not set in GA)
It’s easy to see that Ask Joanne and other brand related keywords are the keywords with the highest search volume, but how else can we make sense of this data.
Here is a word cloud that excludes all phrases with “jo” in them – which covers most variations of my business name and website name. It also excludes all keywords that generated only one visit (to eliminate some of the more fringe words.)
In this word cloud, the size indicates the sum of visits to the site, the color indicates the bounce rate – orange is a higher bounce rate (only one page of site viewed), blue is a lower bounce rate. So, we can see that the phrase What is Yext brought in the greatest number of visitors, and that it had a high bounce rate. However, since it’s a blog post, it doesn’t bother me much that the page had such a high bounce rate. It would be good, however, to track if any outgoing links on that page were clicked (like to my Yext Affiliate link). I should know how to do this.. but with the asynchronous Google code, I actually have to research how to implement it.
You can also look at Word Clouds in other ways:
Data can further be refined, this cloud contains phrases excluding brand name with a bounce rate that is better than the average bounce rate:
Or phrases that brought traffic that visited at least 2 pages per visit: