"I Need to Find That One Chart": How Data Workers Navigate, Summarize and Communicate Analytical Conversations
This is a solid CHI honorable-mention style paper: the novelty is primarily in reframing transcript revisitation as iterative sensemaking and in packaging that idea into a structured design probe. The qualitative evidence is coherent and well aligned with the claims, but the scope remains intentionally narrow.
Axes Lens
Rare contribution shape, typical evidence profile. The point here is not a score. It is to show what kind of claim the paper makes, and whether the evidence pattern is unusual or baseline in this 268 -review set.
Contribution shape
- Knowledge form
- technical knowledge typical · 50/268
- Novelty type
- artifact typical · 20/268
- Abstraction level
- artifact typical · 19/268
- Generalization target
- design family typical · 38/268
- Validation mode
- qualitative study typical · 63/268
Evidence profile
- Evidence strength
- moderate typical · 105/268
- Claim alignment
- strong typical · 231/268
- Overclaim risk
- medium typical · 210/268
Review Summary
This paper’s strongest contribution is not a new model of conversation analysis, but a persuasive redefinition of the problem space. Rather than treating analytical chat revisitation as a matter of scrolling, keyword search, or one-shot summarization, the authors argue that communication and navigation are intertwined in an iterative sensemaking loop. That is a useful CHI move because it shifts attention from transcript retrieval to the workflow of re-entering prior analyses for different audiences and purposes. The design probe makes that argument tangible by layering structure, filtering, multi-level navigation, and detail-on-demand over raw analytical conversations, and by supporting summary authoring with controllable LLM assistance. The qualitative study is appropriately scoped for this kind of contribution: it shows that participants used the system to satisfy recall, reorient, and prioritize needs, and that they employed visual recall, sequential, and abstractive navigation strategies while adding process details and context to summaries. That is enough to support the paper’s descriptive and design implications claims. The main caution is scope. The evidence comes from a small study with a single scenario and a limited participant pool, so the paper should be read as a design and behavioral insight paper rather than a general solution or a causal evaluation of the interface. The authors are also candid that LLM-based extraction of conversational structure may be stochastic or hallucinated, which matters because the probe’s utility depends on the quality of those extracted structures. Overall, this is a credible and well-aimed CHI honorable mention: conceptually sharper than a straightforward interface paper, empirically grounded enough for its claims, but not broad enough to justify stronger generalization or performance claims.
What Changed
Canon before
Prior CHI work on conversational analytics and transcript review largely treats revisitation as scrolling/searching through linear chat logs or as post-hoc summarization. This paper reframes the problem as iterative sensemaking and communication, where navigation and summary authoring are coupled.
Departure from common sense
The paper argues that revisiting analytical conversations is not a simple after-the-fact cleanup task. Instead, communication itself becomes part of sensemaking: people re-enter transcripts to recall, reorient, and prioritize, and those communicative goals drive renewed navigation rather than following it.
Actual novelty
The paper’s main novelty is SyncSense, a design probe that overlays structured conversational elements and affordances on analytical chat so users can filter, navigate at multiple levels, inspect details on demand, and author summaries with controllable LLM assistance.
Evidence
The evidence supports a clear artifact contribution and a qualitative validation of how data workers revisit analytical conversations. The study shows that participants used the probe to satisfy recall, reorientation, and prioritization needs through visual recall, sequential, and abstractive navigation, and to produce summaries with added process details and context. The paper also explicitly frames revisitation and communication as an iterative loop, which is a meaningful conceptual shift from linear transcript review. However, the validation is limited to a small study and does not establish causal effects or broad generalizability.
“1 Design Considerations for the Probe Our research probe serves as shared infrastructure for both navigation and communication tasks. When designing SyncSense , we incorporate affordances drawn from established analytical tools like Jupyter notebooks, spreadsheets, and code repositories, but these features remain unexplored in the context of analytical conversation”
actual novelty · 5 Structured Interface for Revisiting Conversations · confidence 0.97
“ To study behaviors beyond those supported by standard interfaces (i.e., scrolling and keyword search), we develop a design probe that supplements analytical conversations with structured elements and affordances (e.g., filtering, multi-level navigation and detail-on-demand). In a user study ( n = 10), participants used the ”
departure from common sense · 8.2 Limitations and Future Work · confidence 0.98
“ 2023. Tracing and Visualizing Human-ML/AI Collaborative Processes through Artifacts of Data Work. Proceedings of the 2023 CHI Conference on Human Factors in Computing Systems (2023). Digital Library Google Scholar [75] Aayushi Roy, Deepthi Raghunandan, Niklas Elmqvist, and Leilani Battle. 2023. How I Met Your Data Science Team: A Tale of Effective Communication”
limitation · 8.2 Limitations and Future Work · confidence 0.99
“ This reduced engagement heightens the need for tools that support revisitation, navigation, and sensemaking. Revisitation and communication are fundamentally intertwined in analytical practice. Drawing on Pirolli and Card’s sensemaking framework [ 70 ], we recognize that foraging (i”
validation scope · 6 User Study · confidence 0.96
Limits
Method limits
The study uses a small qualitative sample and a single scenario, so it supports rich behavioral insight rather than broad statistical claims. The paper also relies on LLM-based extraction of conversational elements, which the authors note may be affected by stochasticity and hallucinations.
Deployment limits
The probe is most applicable where analytical work is already mediated by conversational systems and where revisitation, summarization, and communication are recurring needs. It is less directly transferable to settings without structured chat logs or without willingness to accept LLM-assisted extraction.
Boundary conditions
Findings are bounded by the specific revisitation task, the participants’ prior analytical conversation, and the interface affordances provided by the probe. The work is strongest for understanding how people navigate and summarize prior analyses, not for proving that the interface improves downstream decision quality.
Position in field
This sits at the intersection of conversational analytics, sensemaking, and transcript revisitation. Its contribution is less a new algorithm than a design-oriented reframing of how analytical conversations should be revisited and communicated, with a prototype that makes that reframing concrete.