8

Most of the sites use data analytics tools like Google analytics. Those analytic tools (the ones I know of) don't provide the analytics for the comments/discussions (I understand it needs to be built into the platform of the site itself). I think it would be helpful for the site administrators if there's a tool that analyzes the comments and provide the useful information for the site administrators, e.g. what is the percentage of the commenters compare to page view.

What are some useful analytics for the site admin in regards to comments? Is there any tool that I can use to analyze the comments?

6

I don't know of a tool that is easily integrated into many system. These types of analytic are built into each platform and the developers of each platform pick out important information for the most common use cases.

I've summarized several key elements from different platforms I've utilized. Depending on how skilled you (or your team) is technically, you may be able to extract metrics from your system even if they don't appear on your dashboard. Additionally, several of these have a nice plugin architecture that allows you to extend functionality, including adding additional metrics to the dashboard.

ArsTechnica provides a nice screenshot of the Discourse Administration panel.

Discouse Dashboard

On this panel, we can see that there are multiple indicators that the developers felt was important. These include new users, new topics, new posts, number of likes, number of flags, bookmarks favorites and emails sent over several different time frames. It also shows the topic topics in the system as well as the top referrers. Discourse also includes various trust levels and nicely shows how many users are in each of those levels.

Wordpress dashboard screenshots are easy enough to find via a Google search.

WordPress dashboard

On this panel, we can see that the focus is more on quantity of actions (number of posts, pages, categories, tags, comments) than a specific break down of events within the system. It also shows the most recent comments in the system as well as recent drafts.

This screenshot is take from one of the PHPBB boards that I administer.

PHPBB Dashboard

Like Wordpress, we can see that PHPBB has a split between the quantity of events (total posts, topics, users, attachments) and averages of events (posts per day, topics per day, users per day). It also provides an quick overview of physical characteristics of the board (size of the database, attachments and avatars).

GitHub contains a few different Dashboards. The first is the project "Pulse". These screenshots are utilizing the Discourse repository because it has a lot of activity and makes for more interesting screenshots.

Github Pulse

From here, we get a very high level view of what has happened. We can see how many contributors we've had in a period of time, the number of changes they have pushed, the number of files and lines changed.

Github also provides a series of "Graphs" to get an idea of how the project is performing. These graphs show the contributions by contributors over time, number of commits per week, frequency of code changes, and the "punch card" or time when commits occur. On smaller repositories, you can also see the forks of the original project.


The above is showing that there isn't a standard set of tools or metrics you can (or should) use to answer all of your questions. Each platform is designed for a certain set of users. That doesn't mean they aren't utilized differently. But the different purposes and uses of each means that each has a different set of "important" metrics.

One thing you may notice is that "views" is not a common metric on the dashboards. This type of information is handled well by Google or alternatives. Combining the two systems is often left as an exercise to the administrator. Merging data from various packages or borrowing metrics requires that you (as the community leader) decide what you want to focus on and determine if you have the information you need to answer the question and then implementing a query to get that data.

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.