AI-assisted qualitative analysis for museum observations

Evidano is an AI-powered qualitative data analysis platform that helps researchers analyze interviews, open-ended surveys, and documents. According to the PLOS ONE study by Wanča et al. (2026), a constructivist Children’s Museum at the National Museum in Prague attracted predominantly preschool visitors and generated measurable gaps between design intent and visitor practice. According to the PLOS ONE study by Wanča et al. (2026), observers collected 720 observation records covering 2, 173 visitors during 16 March to 30 May 2024, producing quantitative measures (interaction, engagement, reactions to labels) and rich field notes that are ideal for AI-enabled qualitative workflows.

Key Takeaways

According to the PLOS ONE study by Wanča et al. (2026) (PLOS ONE), the Children’s Museum opened on 30 June 2023 and an observational protocol (16 March–30 May 2024) produced 720 observations of 2, 173 visitors that reveal a mismatch between the constructivist design and the actual audience. According to the PLOS ONE study by Wanča et al. (2026), "the exhibition is not attracting its intended audience effectively."

720 structured observations were recorded from 16 March to 30 May 2024, covering 2, 173 individual visitors, according to PLOS ONE (Wanča et al., 2026).
Only 23% of child visitors were in the primary target group (ages 6–10) while 72% were younger than the target, according to PLOS ONE (Wanča et al., 2026).
Observers measured an average dwell time of about 7 minutes per observation unit, with U1 and U5 averaging 14 minutes and U2 averaging 2.5 minutes, according to PLOS ONE (Wanča et al., 2026).
The PLOS ONE authors recommend marketing to schools and targeted design adjustments because "learning about the world from a different perspective than provided by the school curriculum" was the exhibition concept (Wanča et al., 2026).

What happened and how the study measured it

What happened: the PLOS ONE observational study by Wanča et al. (2026) evaluated visitor behaviour in the Children’s Museum at the National Museum in Prague using a structured protocol and field notes.

How it was measured: Wanča et al. (2026) deployed 19 trained observers who recorded 720 observation records across five exhibition units between 16 March and 30 May 2024, using 4-point Likert scales for interaction and engagement and sketch maps for movement.

Constraints and design context: the exhibition used a constructivist concept marked by thematic symbols, evocative questions, and a facilitator role, and the study excluded large organised school groups so that the basic unit of analysis was the visiting family, according to Wanča et al. (2026).

Findings snapshot

Date / Period	Metric	Value	Implication
30 June 2023	Exhibition opening	Children’s Museum opened (ChM) at National Museum	Establishes project timeline cited by Wanča et al. (2026)
16 Mar–30 May 2024	Observation sessions	720 observations, 2, 173 visitors	Large dataset for mixed quantitative-qualitative analysis (Wanča et al., 2026)
May 2026 (published 21 Aug 2026)	Age composition	23% target 6–10 years, 72% younger children	Design-target audience mismatch noted by Wanča et al. (2026)
Observation period	Average dwell time per unit	Overall mean ≈ 7 minutes; U1 & U5 ≈ 14 min; U2 ≈ 2.5 min	Unit design affects engagement and time-on-task (Wanča et al., 2026)
Slots/days	Facilitator engagement	Facilitators mean score 0.27 (SD 0.63); higher in U5	Facilitator workload and slot scheduling influence mediation (Wanča et al., 2026)

Implications for museum researchers and UX teams

Implication summary: the PLOS ONE evidence suggests researchers should combine structured observation with scalable qualitative synthesis to surface audience-design mismatches quickly.

For evaluators: Wanča et al. (2026) show that a 2.5 month observational protocol with 720 records can quantify interaction types (focus, communication, creativity) and reveal unit-level effects; AI-enabled coding accelerates replicating this at scale.

For designers and educators: Wanča et al. (2026) found that younger visitors rely more on adult guidance and less on evocative questions and labels, which implies designers should test label visibility and facilitator scripts with the actual audience before launch.

How Evidano helps: from noisy observations to actionable themes

Problem: large mixed-format data that take months to code

Answer: Evidano is an AI-powered qualitative data analysis platform that helps researchers analyze interviews, open-ended surveys, and documents.

Evidano feature mapping: Wanča et al. (2026) generated 720 structured forms plus open field notes, and Evidano can ingest mixed formats (transcripts, observation spreadsheets, and field notes) to produce thematic and frequency analyses in hours rather than weeks.

Problem: hand-coding engagement and interaction dimensions

Answer: Evidano automates initial coding and surfaces reviewer-approved themes.

Evidano capability: Evidano supports hierarchical coding and cross-segment comparisons (e.g., preschool vs. junior-primary), which mirrors the unit-by-age comparisons used by Wanča et al. (2026), enabling fast replication and hypothesis testing.

Problem: spoken field notes and facilitator interviews

Answer: Evidano provides speech-to-text and PII redaction to convert audio notes into analysis-ready text.

Relevant feature: use Evidano's transcription tools (Speech-to-text) to transcribe facilitator debriefs and visitor interviews, then run thematic extraction across transcript and protocol data.

Problem: iterating exhibit changes and measuring impact

Answer: Evidano creates repeatable dashboards and AI chat over your documents so teams can ask questions like "Which units increased creativity scores after redesign? " and get immediate evidence.

Try features: see how thematic tagging, co-occurrence networks and cross-segment frequency counts turn the raw observations Wanča et al. (2026) collected into testable design decisions; learn more on Evidano's features.

FAQ: ai qualitative analysis museum observations

How can AI speed up coding of structured observation protocols like the PLOS ONE study?

Direct answer: AI can pre-code observation text and scalar annotations, then present human reviewers with suggested codes for rapid validation.

Supporting detail: Wanča et al. (2026) combined Likert scales with open notes; AI-assisted pipelines ingest both numeric and free-text fields to cluster themes, surface high-frequency phrases, and produce codebooks that match human reliability in hours rather than weeks.

Can AI preserve the nuance in observers' field notes while scaling analysis?

Direct answer: yes, when AI models are tuned for qualitative nuance and reviewed by domain experts.

Supporting detail: The PLOS ONE dataset includes contextual sketches and diaries; AI thematic models can flag candidate quotations and preserve sentence-level context while enabling researchers to filter by age cohort, unit, or slot for targeted interpretation.

How should museums use AI outputs to change exhibits responsibly?

Direct answer: use AI outputs as evidence to run small, testable interventions and measure the same metrics before broad rollout.

Supporting detail: Wanča et al. (2026) recommend targeted redesign and school outreach because 23% of child visitors matched the intended age; museums should pilot label placement or facilitator scripts, measure dwell and engagement, and iterate using the same observation protocol.

Conclusion & Next Steps

Summary: the PLOS ONE observational dataset (Wanča et al., 2026) demonstrates how structured protocol data plus field notes reveal audience-design mismatches and unit-level engagement patterns that are ideal for AI-enabled qualitative analysis.

Actionable next step: museum researchers should combine the PLOS ONE protocol approach with AI-assisted transcription and thematic analysis to shorten the evidence-to-decision loop.

If you want to try this workflow, upload observation protocols, transcripts, and field notes to Evidano and test automated coding and cross-segment analytics; for details, start at Evidano's features page (Evidano features).

Ready to test your own dataset? Try Evidano for free.

Topics

ai qualitative analysis museum observations
ai-assisted qualitative analysis
qualitative analysis of museum visitors
thematic analysis museum observations
ai transcription for ethnography

Keep reading

Browse all articles