.

KONDA: User Feedback Survey

Research Data Management (RDM) is increasingly important for making research data organized, reusable, and easier to share. We've built an app to help researchers turn their datasets and related documents into structured knowledge graphs aligned with ontologies using large language models (LLMs).

This survey will help us evaluate the app’s usability, usefulness, and impact on research workflows.


What You'll Do

1. Answer a few questions about yourself

2. Learn how the app works

3. Use the app with a provided sample dataset

4. (Optional) Try it with your own data

5. Give feedback on your experience


The aim of this study is to evaluate the developed app in terms of its usability, its applicability for research data and its potential added value for research data management. The central question is: To what extent can researchers use the tool to transfer their datasets into knowledge graphs more easily and effectively.

Secured
KONDA: Usability Feedback Survey

If you consent to participate, you will be asked a series of questions which you should answer as truthfully as possible. Your answers are anonymous and will be stored securely. The survey will take approximately 30 minutes to complete.

Participation in this study is entirely voluntary. You are free to stop or withdraw at any time without providing a reason and without any repercussions.

By continuing, you acknowledge that you have read and understood the above information and that you consent to participate in this study.

1

If you have read and understood the above, please confirm your consent to this study.

Wählen Sie eine Antwort
KONDA: Usability Feedback Survey
2

What is your research area?

e.g., computer science, mathematics, physics, chemistry, biology, engineering, ...
3

What is your current position?

Pick the most applicable one, or specify otherwise
4

Experience with research data management (RDM)

How would you rate your experience with research data management (e.g., organizing, storing, sharing research data?)
5

Technical expertise

How would you rate your general technical skills (e.g., programming, working with computers and data)?
6

Familiarity with ontologies

How familiar are you with the concepts of ontologies?
7

Familiarity with knowledge graphs

How familiar are you with the concepts of knowledge graphs?
8

Commonly used data formats

Which data formats do you typically use in your research?
9

Commonly used research data types

What types of research data do you primarily work with?
10

Openness to new tools

How open are you to try new tools for managing or processing your research data?
KONDA: Usability Feedback Survey
11

Understanding Ontologies

When working with research data, it's important not just to name things, but to ensure they’re understood the same way by everyone - including machines. That’s where ontologies come in. An ontology is a structured vocabulary. It defines what kinds of things exist (e.g., Person, Paper, University) and how they relate to each other (e.g., authoredBy, affiliatedWith). Each concept is given a clear meaning and a unique identifier (URI) so that systems and people can speak the same "data language". This diagram shows an ontology as a general model: It defines the types of things and the possible relationships - but without any real-world data filled in. Think of it as a blueprint providing the vocabulary: it shows what kind of information can be captured and how it can be connected.
question image
12

Knowledge Graphs

A knowledge graph brings an ontology to life. While an ontology defines the types of entities (like Person, Paper, or University) and their possible relationships (like authoredBy or affiliatedWith), a knowledge graph uses that structure and fills it with real-world data. It represents entities as nodes and their relationships as edges, forming a network of semantically connected information. For example: - Researcher B is affiliated with University A - Paper D is authored by Researcher B and published in Journal E This graph-based model captures not just isolated facts, but the context and meaning behind them. The result is a richly connected, machine-readable representation of a domain that supports complex queries, data integration, and deeper insights.
question image
KONDA: Usability Feedback Survey
13

KONDA - Overview of the App Workflow

Now we’ll introduce KONDA - a prototype developed as part of a master's thesis to demonstrate that datasets can be transformed into knowledge graphs with minimal prior knowledge of ontologies or knowledge graph concepts. Note that while the current output is functional, it can be further refined to provide more accurate and customized results for your applications. This diagram shows the main steps of the app for transforming your research data into a knowledge graph: 1. Upload Files - Add your dataset and context documents, and select your research domain. 2. Select Ontologies - Search for relevant ontologies using keywords from your research. 3. Entity Recognition - The app suggests key entities from your data. 4. Relation Extraction - It also identifies potential relationships between those entities. 5. Ontology Annotation - The entities are mapped to standard ontology terms. -> In steps 3 to 5, the user needs to review and verify the suggestions to ensure they reflect the research correctly. 6. Explore Graph - View and export the final knowledge graph built from your input.
question image

You’re now going to test the app using a predefined dataset. Please follow the steps below carefully. This part is essential to understanding how the app works and how usable it is for your research workflow.

Prerequisites – Before You Begin

Make sure you complete the following setup steps and then come back to the survey:


1. Download the dataset and readme from the links provided below (please copy paste these links into your URL bar and save the data):

https://rwth-aachen.sciebo.de/s/TTdgbjdGV68rzVa

https://rwth-aachen.sciebo.de/s/ZSj2qwc3gtSbxe1

(A dataset about notes on cultivating edible plants from a community)


2. Open the app in a new browser tab. (Note: The app must be accessed from within the RWTH network.)

https://cloud22.dbis.rwth-aachen.de/


3. Register an account by choosing a username and password.


4. Use the following invite code to complete your registration:

KONDA-USABILITY-FEEDBACK-SURVEY


5. Come back to the survey

14

Prerequisites

Select one answer

Step-by-Step: Try the App with the Provided Dataset

Hint: In case you are interested in more details, there is a Help button in the bottom-right corner, with some more explainations of each step (help texts appear at the top of the page).


Get to know the dataset:

1. Open the readme file and the csv file (inside the zip) and briefly familiarize yourself with the dataset


Screen 1:

1. Go to the tool

2. Upload the dataset zip file and the readme you downloaded earlier.

3. Select the domain: Gardening

3. Enter this description in the provided text field:

 -> "A data log about growing edible plants in different locations"

4. Click Submit


Screen 2:

1. Keep the Ontology Settings at their default values, except:

  Enter 5 in the "Number of Entities to Extract" field (optionally also run it with 15 and 20).

  KONDA now tries to find the 5 best Named Entities and Categories that can describe the dataset. The details get increased if we increase the number of named entities to extract.

2. Enter "plant" in the field "Search for Ontologies" and select the Ontology "cgo".

3. Click Submit


Screen 3:

KONDA has now found 5 named entities and suggested broader categories for each one. 

1. Click on one category to see other suggestions.

2. Make sure that the suggestions make sense for the entity - change them if they do not.

  All values here are plain text and you can also write your own entities or categories.

  KONDA then tries to find fitting ontology terms later on.

3. Click Submit


Screen 4:

KONDA has now found 5 relations between the named entities and categories.

1. Click on one relation to see other suggestions.

2. Make sure that the suggestions make sense as a relation - change them if they do not 

  Again the relations here are just plain text which you can edit.

3. Click Submit


Screen 5:

KONDA is now trying to map your plain text entities, categories and relations to actual ontology terms.

You can see the entities and categories in blue, while the relations are displayed in red.

1. Again we need to make sure that all mappings are correct here

2. If one does not fit, click on the ontology term input field and try to search for a more fitting term

3. Find a blue match that exactly fits the ontology term (e.g. "Crop" matched to "Crop") and select the blue checkbox to replace the custom label with the ontology term.

  (remember the name of the ontology term for the next screen)

4. Click Submit


Screen 6:

You have now created a Knowledge Graph.

1. Explore the graph displayed at the top.

- At the center, you'll see your dataset (yellow node).

- It's connected to your context files (purple nodes) via a "has context file" link.

- The dataset is also connected to the main concepts (blue nodes, named entities) using "has concept" links.

- Each concept may be connected to a broader category (orange node) with a "has broader" link.

- Concepts (blue) and categories (orange) can also be linked to ontology terms (green nodes) using "has related match".

2. Click on a blue node and notice the custom URI that was created.

3. Notice how blue and orange nodes are connected to green nodes.

4. Find the node that you have changed in the step before

5. Notice that this node has no "has related match" connection. Instead the label and URI was directly taken from the ontology term and replaced the custom label and URI.

6. Scroll to the bottom of the screen and export your knowledge graph, by selecting Turtle as the format.

15

App Execution with Provided Dataset

Select one answer
16

If you had problems, please describe them below

Optional: Try with Your Own Data

If you'd like, you can now repeat the same process using your own research data to see how the app performs in a real-world scenario. This step is optional.

17

Did you use the app with your own Dataset?

Select one answer
KONDA: Usability Feedback Survey
18

App Usability

19

Additional Usability Feedback

In the following questions, please answer as truthfully as possible.
20

Overall usability rating

How would you rate the overall usability of the app? 1 = very poor → 10 = excellent
KONDA: Usability Feedback Survey
21

Functionality and Perceived Impact

Now will follow a series of questions regarding the perceived impact of the app.
22

(Optional) What would make the app more useful in your daily research work?

23

Prior use of similar tools

Have you used any tools before for converting research data into knowledge graphs or working with ontologies?
24

If yes, please specify (optional)
25

Feedback (optional)

What did you find particularly easy or difficult to use? What improvements to current functionality would you like to see?
KONDA: Usability Feedback Survey

You made it through the end of the survey!



Thank you very much for your participation.