🚧 Dembrane Docs is under construction - Things will change
ECHO
Analysis

Your First Analysis

You have set up your project and conducted your first conversations with ECHO 👏 With the analysis functions, you can now start analyzing the collected input and extracting insights.

The analysis techniques discussed here:

  • Ask question: Query your collected data in a flexible way.
  • Report: Generate a report that tells which questions the collected conversations answer.
  • Library: Generate a cluster overview of themes that emerge in your participants' conversations.

Using the Ask question function

The Chat function enables detailed analysis of specific conversations in a conversational way, allowing you to extract meaningful insights from your data.

Selecting Context

  • If you have clicked on a conversation, and then click on "Ask question," this conversation will automatically be selected as context.
  • If you click on "Ask question" without currently viewing a conversation, it will take you to the chat interface without any specific added context. You will need to add context yourself in this case.
  • You can always add more context to your chat by checking the boxes next to individual conversations.

Best Practices

  • Start with broad questions. For example: What is being discussed in these conversations?
  • After getting a high-level overview, you can zoom in. Here it helps to ask clear and specific questions.
         - Use follow-up questions for clarity
         - Break complex questions into smaller parts
  • Document important insights

Example Questions

Below you'll find some examples:

  • What arguments are presented when discussing X
  • What seem to be the biggest differences in values between recording X and recording Y
  • What agreements were made in this meeting? What's still missing to make these agreements SMART?
  • Can you give me the quotes on which you based [statement in the chat]?

Creating a Report

The report function makes sharing general trends in your collected conversations easier than ever. How this function works is that behind the scenes the question is asked: what questions do these conversations answer? The result can be found in the report. You cannot edit this report yourself. If this is desired, you can always contact Dembrane.

In the portal editor you can enable participants to opt-in to receive updates of the report. After completing the recording they will have the option to leave their email adress. When a report is published of the project the link will be sent to those participants directly.

The Library Function

Libraries are a great way to get a broad overview and a feel for what participants talked about, clustered around questions or topics of your choice!

Creating a library

  1. Click on the "Library" button in the sidebar of your dashboard. This takes you to your project library.
  2. Initially, your library will be empty, as indicated by the blue status message at the top of the page. Here you can also see how many conversations are waiting to be processed
  3. Now click on "Create Library". By default, the library will create two so-called "Views" - Topics and Sentiment. Views are the subjects, questions, themes of your choice that the AI will use as a basis for clustering the conversation quotes

  1. Click on "Create Library" and wait for the processing to complete. This may take some time
  2. Once the library has been generated, you can start exploring the two default views - Topics and Sentiment - which should now have so-called "Aspects" linked to them. Aspects are AI-generated clusters around the question or query of the view. Click on one of them to start exploring how conversation quotes were clustered around the aspect.
  3. After creating your first library, you're ready to create your first unique views!

Note: You might notice in the library that some cluster themes have 100s of quotes, while others have few or even zero. Often the clusters with many quotes are the 'open doors', clusters that you can expect to appear beforehand. The quotes with fewer citations are the ones that the AI identifies as relevant, and can be interesting paths to explore with your participants.

Creating Additional Views

  1. In your project library, click on "Create View"
  2. Enter your query: Here you can ask your research question
  3. Additional context: This is about what you want the AI to pay particular attention to, or for example how many clusters you want (such as a top 10 list)
  4. When you click on "Create View", the AI will begin gathering relevant quotes

Exploring Views

There are multiple ways to explore a view.

  1. By clicking on the expansion icon in the top right corner of your view, you'll see a more detailed overview of the view
  2. You can click on any of the aspects at any time to begin exploring how conversation quotes were clustered around the aspect
  3. For each quote, you see the conversation source linked to it. When you click on it, you go directly to the "Analysis" tab of the conversation, which not only contains all extracted quotes but also AI-generated Insights

Now you are well equipped to create a project, share it with your participants, and start collecting data.

Need Help?

Contact Jules if you need help! jules@dembrane.com