Did you collect a lot of posts in your project? Great! But now the real work begins: analyzing all that text-based input and transforming it into actionable insights worth sharing. Our insights tool helps you do just that. Follow the steps below to make sure the time spent on processing your posts is minimal while getting out maximum value.

Note: The actions taken in our insights tool won't impact the rest of your platform and won't be visible to participants.

Select the project you want to analyze and name your new insights.

You can create as many insights as you need starting from the same project data.

After creating your insight, you'll land on the exploration page. Here, you can browse and filter your posts by interacting with the keyword visualization, by filtering on tags (cf below), or searching. By exploring your posts here, you get a good understanding of what people are talking about. If you want to fully analyze your project and draw conclusions, this also proves to be a great starting point for all steps below.

The visualization shown is a network visualization. It consists of the important and recurring keywords automatically detected in your project's posts.

The big circles represent groups of keywords that often occur together, and the most important of those keywords are shown inside those big circles.

Clicking such a big circle opens it up to show the individual keywords that are linked with it. Clicking a keyword adds it as a filter to the list of posts on the right-hand side.

Some practicalities:

When processing and analyzing the collected posts, an important step is to group the posts that belong together in a way that makes sense for your project, your needs, and your context. That's where 'tags' come in. You can define them as you see fit and create as many as you need.

Once the posts are tagged, it will be much easier to dive in, summarise and draw conclusions to base your decisions on.

Your custom project tags are used as default tags, prefilled with the posts that were manually assigned to that tag by the author. Don't need them as tags? Simply click 'Reset tags,' and you can start creating tags from scratch.

If you have a good sense of those tags (e.g. by browsing the first X input items), you can manually add your tags in the 'Add tag' input box.

What form could tags take?

Now that you've defined your tags, it's time to group all collected posts into those tags. Posts can be assigned to zero, one, or multiple tags.

There are again two ways of assigning posts to tags:

[Premium customers only] The algorithms used to tag the posts will never be 100% 'correct'. Therefore, when accuracy is required, you can use the optimized annotation flow to go through the tagging results and manually correct and complement them.

There's a clear visual distinction between approved or manually added tags on one hand and those that got added automatically without approval. The posts in each tag can also easily be sorted to bring the non-approved items on top.

A post with a manually added or approved tag (left, 'Health and welfare') and an automatically added one that requires approval (right, 'garden')

Once more, there are two ways to complement and correct the categorization:

Posts that get added to your project after creating your insights will be assigned to the 'Recently posted' section, where you can assign them to your existing tag structure.

Once you've completed the previous steps, click 'Done' to return to the exploration view.

Here, you can easily filter on each of your tags and discover the participant input that comes with it. This will allow you to understand each tag well and draw your project's conclusions.

Once done, you can now export the tagged input to Excel.

The export file will contain a column for each tag, with 'approved' indicating that the tag has been manually added to or approved for the post of that row and an empty cell indicating that the post is not linked to that tag. Automatically assigned tags that haven't been approved have a 'suggested' value.

The export can be used as a starting point to create visuals for your report, or it can serve as an internal tool to browse your posts in a more structured way.

The first thing you will see once you have created Insights is what we call the "Exploratory View." On this view, you can browse and discover the participants' inputs by

You can use Exploratory View for two primary actions;

Using in combination with Tags to search for specific inputs

Combining Tags that are already available with the keywords allows you to look for particular inputs that apply to both or have a closer look at the tags you have already created.

For example, if you pick the Tag "Arts & Community" and keyword "Youth"; the result will filter the list of posts that are related to the keyword "Youth" under the "Arts & Community" Tag. This new list of inputs can be saved as a new Tag.

Using Keywords to categorize your inputs into Tags efficiently

If you have not used the default Tags or find the currently available Tags on the platform irrelevant for your analysis, you can use the exploratory view to search and create tags quickly.

You can start by exploring the more prominent keywords which represent more frequently discussed concepts. Then, by picking a keyword or similar keywords, you can quickly categorize them into Tags. With this action, you can organize hundreds of inputs within minutes.

For example, you can select keywords such as "waste", "recycling", and name the Tag as "Waste Management". All the inputs that are relevant to "waste" and "recycling" will now be categorized under "Waste Management".

The Dashboard View is where you explore, analyze, categorize and manage the inputs. You can quickly click on the inputs to have a detailed look at individual inputs and toggle between inputs with UP or DOWN arrow keys on your keyboard.

Creating Tags to categorize efficiently

Detecting inputs under a Tag

If you were to create manual Tags - it is advisable to give them descriptive labels at the start. Our algorithm uses the labels as the "search word" to look for relevant inputs in the project and suggest them to you. You can rename it to a label that is fit for reporting purposes at the end.

For example; Using "Car Accidents" as a tag name would get you more accurate automated suggestions than "Mobility issues".

How is the size of the keywords determined in the visualization?

Size represents the simple frequency of the keyword occurrence within your project. The more frequently the particular keyword or its variation, i.e., tree and trees appear in the project, the bigger the keyword appears in the visualization.

How are keywords selected and clustered together in color?

We first determine the keywords that appear in the same input. If the keywords appear together in the same input, we consider it a "connection" between them. Then we use the Louvain method, a mathematical model in which the number of connections between two "keywords" are maximized and clustered into an optimum number of color groupings. Optimizing the number of connections theoretically results in the best possible clusters of the keywords.

How is the proximity or the connection line between the keywords are considered?

The stronger the connection between the two keywords, i.e., the amount of co-occurrence, the closer the two keywords appear and the darker the line connections between them. Then we use the forced-directed layout algorithm, which positions the network in a two-dimensional, easy-to-read way.

Tip: In some instances, you may see a cluster of keywords appearing by itself apart from the rest of the clusters. This signals that this issue is discussed separately and not relevant to the other ideas.

Understanding the automated analytics

We use Natural Language Processing (NLP) to give you automated suggestions of inputs that are relevant to your Tags.

Here is how it works:

How do you analyze which inputs are relevant to the specific Tag?

We use a Natural Language Inference (NLI) model based on BERT, which determines the meaning and context of the Tag and determines whether the input provided is semantically related to the Tag. If the model determines a relevant relation between the Tag and the input with high confidence, we suggest the Tag.

For example, the language model would determine the word "Nature" is highly relevant to the input "I wish there are more trees in the city", and suggest this input for the Tag.

How many languages can Insights work on, and how accurate is it?

Our Insights feature currently works in 16 languages, including; English, Dutch, German, French, Spanish, Portuguese, and Danish. We are constantly working on adding support for more languages. It is considered one of the most accurate models and can detect the contexts of the keywords (e.g., the relation between trees and green). However, as with any computational model, there is a certain degree of false positives & negatives and requires human supervision.

Need more help or support? Contact our Support Team via the chat bubble.