Short answer
When designing for expert users handling complex networks, move from 'showing data' to 'providing tools for data interrogation' by including adjustable thresholds and source transparency.
- Field
- User-Centred Design
- Source
- Nucleic Acids Research (2020)
- Method
- Iterative software development and system architecture update
- Sample
- 14,000+ organisms indexed
- Evidence
- Strong effect
Users perceive high-density data as more reliable when the system provides transparent confidence scores and allows for manual manipulation of the visual hierarchy. This user-centred design research insight is drawn from a 2020 study published in Nucleic Acids Research. Using Iterative software development and system architecture update with 14,000+ organisms indexed, researchers explored how this design variable affects real-world outcomes. The key design takeaway: When designing for expert users handling complex networks, move from 'showing data' to 'providing tools for data interrogation' by including adjustable thresholds and source transparency.
Automated evidence scoring and visual network customization increase trust in complex data visualizations
Users perceive high-density data as more reliable when the system provides transparent confidence scores and allows for manual manipulation of the visual hierarchy.
Nucleic Acids Research · 2020
Key Findings
- 01Multi-source evidence scoring reduces the perceived risk of false positives in automated data mining
- 02Physical interaction 'scoring modes' allow users to toggle between functional and structural views, improving task relevance
- 03Automated enrichment detection identifies biases in user-uploaded data, acting as a built-in quality control
Application
Design takeaway
When designing for expert users handling complex networks, move from 'showing data' to 'providing tools for data interrogation' by including adjustable thresholds and source transparency.
How to apply
In a dashboard for financial risk or supply chain logistics, add a 'Confidence Threshold' slider that hides or shows connections between entities based on the system's certainty score.
Project actions
- 01Use color-coding to represent the strength of a connection between two items
- 02Always provide a way for the user to click a link and see the 'source' of the information
- 03Allow users to drag and reorganize elements in a map to help them make sense of it
Method & Evidence
Strengths & Limitations
Limitations
This approach requires a backend that can actually calculate 'confidence,' which might be hard for a simple student prototype.
Think critically
Does showing a 'confidence score' make a user more critical of the data, or does it make them blindly trust the numbers the computer gives them?
Design Principles
"Transparency breeds trust in automated systems."
In data-heavy environments, cognitive load is high; providing a 'confidence score' for system-generated connections helps users filter noise. By allowing users to customize and extend networks, the interface shifts from a static display to an active workspace, fostering a sense of agency over complex information.
What This Means for Your Design
When you have a lot of messy data, giving users a way to filter it by 'how sure the computer is' makes the tool much more useful and trustworthy.
How to use in your project
- 1.Cite this when justifying the use of filters or 'advanced settings' in a data-heavy interface
- 2.Reference the 'scoring mode' concept when explaining why your design allows users to switch between different views of the same data
Add to My Project
Quick Cite
Paragraph starter
According to research on the STRING database (Szklarczyk et al., 2020), providing users with customizable networks and evidence-based scoring improves the functional characterization of complex data sets.
Source
Nucleic Acids Research
The STRING database in 2021: customizable protein–protein networks, and functional characterization of user-uploaded gene/measurement sets
journal · 2020
View sourceQuestions About This Research
- What does the research say about automated evidence scoring and visual network customization increase trust in complex data visualizations?
- When designing for expert users handling complex networks, move from 'showing data' to 'providing tools for data interrogation' by including adjustable thresholds and source transparency. Evidence: Nucleic Acids Research (2020).
- Why does "Automated evidence scoring and visual network customization increase trust in complex data visualizations" matter for design?
- In data-heavy environments, cognitive load is high; providing a 'confidence score' for system-generated connections helps users filter noise. By allowing users to customize and extend networks, the interface shifts from a static display to an active workspace, fostering a sense of agency over complex information.
- How can designers apply this research?
- When designing for expert users handling complex networks, move from 'showing data' to 'providing tools for data interrogation' by including adjustable thresholds and source transparency.
- What were the main findings?
- Multi-source evidence scoring reduces the perceived risk of false positives in automated data mining. Physical interaction 'scoring modes' allow users to toggle between functional and structural views, improving task relevance. Automated enrichment detection identifies biases in user-uploaded data, acting as a built-in quality control
- What research method was used?
- Iterative software development and system architecture update with 14,000+ organisms indexed.
- How strong is the evidence?
- Evidence strength is rated Strong effect, based on a 2020 journal from Nucleic Acids Research.
- What should I do differently in my next project?
- In a dashboard for financial risk or supply chain logistics, add a 'Confidence Threshold' slider that hides or shows connections between entities based on the system's certainty score.
- What are the limitations?
- The complexity of the interface may create a steep learning curve for non-experts; high-density networks can still lead to 'hairball' visualizations despite filtering.