About a year and a half ago, we started realizing that GovLoop would be facing a significant problem over time. Members were producing hundreds of pieces of content each month in the form of blogs and and discussions. While that’s a great problem to have as it signifies a vibrant, active community, it also meant that we needed to start making it easier to categorize, discover and share that content more readily. Here was our process to address this problem:
1. Figure out what tags people are using. First, we embarked on a project to learn how people were tagging content. GovLoop member Sam Allgood faithfully helped us to search the site and update those lists every week for well over year as a volunteer (thanks, Sam!). and to make those tags available to others who might want to more quickly find information that’s relevant to them. You can find our tag lists over on Socrata:
2. Standardize tagging based on most popular topics.The next step was to sort through those tags and determine what were the most popular. That’s where Sara Cohen stepped in to help.For the past several months, we have been working to enhance the site by developing new tagging procedures. We settled on 10 categories and the GovLoop team started coming behind each blogger to make sure all content was grouped in one of the following categories:
3. We made it easier for members to tag content in this way. You can read more about that part of the project here. Using the new drop down menus (versus adding your own tags) will help to ensure that other Govloop members will find your posts, even when you call it TO-MAY-TO, and they call it TO-MAH-TOE.
4. We reviewed the more than 18,000 blog posts written by Govloop members since the site began, and created a list of the most talked about categories. We then developed a list of sub-tags, or key words, often used in blog posts about each category. For example, in a blog post about technology, one might use the words “social media,” “data,” “software as a service,” etc. We took those sub-tags and included them with each of the categories (a total of more than 200 sub-tags [268 to be exact!] that will continue to grow).
5. We then developed an algorithm that went through all of the blogs and analyzed them for the sub-tags, and then assigned a category (or multiple categories) for each blog post in the form of new tags added to the existing tags. Rest assured, when you search for “technology” now, the blogs that talk about technology will show up in your search results, even if they don’t include the actual tag.
Long story short (and multiple lines of code later), the content on Govloop is now more searchable (and findable) than ever. Go ahead and try it!