David Simpson

Want to know who are the most active users on your Confluence wiki this week?

This wiki markup snippet should help. It tabulates:

• Overall activity
• Page creations
• Page Updates
• Page views
• Page Popularity

h1. This Week's Wiki User Activity Charts
 
h2. Overall
 
{topusers:spaces=@all|timespan=1w|period=weekly|events=create,update,view|display=icon,title,count}
 
h2. Breakdown
<code>
|| Page Creation || Page Updates || Page Views || Page Popularity ||
| {topusers:spaces=@all|timespan=1w|period=weekly|events=create|display=icon,title,count} | {topusers:spaces=@all|timespan=1w|period=weekly|events=update|display=icon,title,count} | {topusers:spaces=@all|timespan=1w|period=weekly|events=view|display=icon,title,count} | {popular:spaces=@all|timespan=1w|period=weekly|events=view|display=icon,title,count} |
</code>

I love Confluence. It’s simple. It’s easy to use. It’s hierarchical, so you can build a nice structure.

But…

I hate Confluence. I can’t find anything. Ever. It’s hierarchical, but which branch in the hierarchy is my content located in? Damnit, the Lucene powered search doesn’t seem to be able to find what I’m searching for.

My main gripe with Confluence is that by default it uses breadcrumb trails as primary navigation contrary to breadcrumbs best practices. They are no replacement for the real deal.

Breadcrumbs are secondary navigation designed as an alternative to the main navigational scheme. As such they give no context to relate the current page to others at the same level in the hierarchy. Breadcrumbs cannot aid in navigation within a level. So using breadcrumbs as primary navigation leads to a lack of awareness of the structure of the website/wiki space and thus confusing.

With this in mind, I decided to add a consistent primary navigational scheme to our global Confluence template. As Atlassian have in the past been known to change the templating scheme a little on upgrade to newer versions of Confluence, it is important to limit the changes to be as minimal as possible to limit ongoing templating maintenance issues.

You’ve chosen Confluence as your Enterprise wiki and got it installed. People are using it. But how many people?

You could have a look at the Global Activity page on Confluence, but this might confuse matters.

You’ll get to see graphs, but what do they really mean? The Global Activity statistics are at best very basic.

Better still, install Google Analytics and you’ll have a much more accurate idea of what’s going on. Best of all, it’s really easy to install.

If someone using a desktop Twitter client (such as TweetDeck or Twirl) clicks on a link I’ve posted to my blog, the Google Analytics referer information will be blank. It registers as direct traffic to the website.

This is fine when linking to other people’s websites — I’m not worried about them — but I’d like to know where that traffic to my websites comes from.

I generally shorten URLs with a service like bit.ly (which does provide basic statistics on click throughs). The trouble here is that you still miss the referrer information on your Google Analytics account.

How can we capture that the source of the link was a Twitter post?

The Analytics Talk blog post — Twitter and Google Analytics: What to Track — gives a good solution to this problem.

Put simply, before you shorten the URL, add the following GA campaign request parameters to the end:
?utm_campaign=blog&utm_source=twitter&utm_medium=micro-blog

e.g.
http://example.com/
becomes
http://example.com/?utm_campaign=blog&utm_source=twitter&utm_medium=micro-blog

Now you can track referrers from Twitter clients within Google Analytics using
Traffic Sources | Campaigns. Nice.

Check out Twitter and Google Analytics: What to Track by Justin Cutroni for the details.

Aside: Justin’s Google Analytics Short Cut PDF book ($10 from O’Reilly) is a great read, but it was written in 2007 before the last GA update. I hope Justin gets the time to release an updated version soon.

If your website contains Flash movies that link offsite or to microsites, you may wish to register the offsite link using WebTrends. This article provides an introduction. It assumes a little knowledge of ActionScript, Javascript as well as WebTrends.

This example uses a intermediary javascript function that sits between the Flash movies and the WebTrends code.

Benefits

• It’s much easier to debug if the WebTrends javascript is not called within the compiled Flash movie
• The WebTrends tracking can be changed later without recompiling the Flash movie or even speaking to your Flash developer

The slides from a presentation I’m giving today at the European Luminis User Group (ELUG) meeting at the University of Greenwich.

To find out which of the UK’s national newspapers has the best readability, I chose a news story at random that was recently covered in each of the newspapers.

I tested the readability of each newspaper’s article using my own online readability testing tool.

I’ve just released the first version of The Readability Testing Tool.   It takes either a supplied URI, or directly inputted text.

Indices measured

• Flesch Kincaid Reading Ease
• Flesch Kincaid Grade Level
• Gunning Fog Score
• Coleman Liau Index
• Automated Readability Index.

Flesch Kincaid Reading Ease is a scale of 0-100. A high score means the text is easier to read. Low scores suggest the text is complicated to understand. All the other indices are based on the US school grade level system.

Features

I’ve added features that aren’t available anywhere else.

1. Readability testing by referer.
   Allows you to add a link to the site on your blog, so that you can rapidly view the readability scores
2. Readability testing a specific part of the web page.
   If the article text in your blog is contained in e.g. div#content, this can be used to test only the article text.  This means that the result is not affected by the navigational elements.

This is a step by step guide for building a find nearest/store locator for free in PHP just like Postcode Anywhere.

It requires PHP and a mySQL database. That’s all.

You don’t need Google Maps, but it helped me populating the database, so I’ve included it for completeness. The example I’m using is UK based, but provided Google have a postcode/zipcode to lat/long convertor for your country, it should work nicely.

Since I last added Google Analytics to a website, Google have changed from the old urchin.js file to the shiny new ga.js file. This means that I really ought to start using ga.js now.

The University of Nottingham uses Sungard’s Luminis portal system which produces some rather nasty looking URLs. Web analytics for this is rather a handful – bad URLs, multiple features per page (each of which we’d like to track) and often tracking would be nice across sub-domains.

With that in mind, and Google’s Tracking Code Migration Guide close at hand, I’ve written jGoogleAnalytics.js to allow me to track:

• clicks events
• form submit events
• cross subdomain
• cross domain (e.g. for eCommerce payment gateways hosted externally)
• new organic search engines
• all the features of Jason Huck’s GA jQuery integration

Download jGoogleAnalytics.js