from delay state j 1 LGA 30 NY j 2 LGA 40 NY j 3 JFK 50 NY j 4 IAH 60 TX Provenance for Interactive Visualizations Fotis Psallidas [email protected] Eugene Wu [email protected] Introduction Provenance For Interactive Visualizations Lineage Primer Provenance For Interactive Applications Interactive Selections • Item selection • Group selection • Range selection • Generalized selections airlines airports ontime states View V 1 backward_trace() Logic over Selections • Tooltips • Details-On-Demand • Semantic Zooming airlines airports ontime states View V 1 Details V 2 V 1 V 2 Idaho Montana SouthDakota NorthDakota Wyoming avg_adelay avg_adelay Multi-View Linking • Linked Brushing • Crossfilter SQL(backward_trace()) selective_refresh(backward_trace()) γ "#$#%,$'((*%+$,) (Airports⨝Flights) name = from name state a 1 LGA NY a 2 JFK NY a 3 IAH TX from delay f 1 LGA 30 f 2 LGA 40 f 3 JFK 50 f 4 IAH 60 airlines airports ontime states V 1 Search Profile Tool Vis App External Apps Data Store [VLDB18] Smoke: Fine-Grained Lineage At Interactive Speed [SIGMOD18] A Deep breath of Data-Intensive Lineage Applications [HILDA18] Provenance for Interactive Visualizations [CIDR17] Combining Design and Performance in a DVMS Want to Know More? Example: Multi-Application Linking Provenance management systems can provide • core functionality across application domains and • novel functionality to extend the space of interactive applications state avg(delay) O 1 NY 43.33 O 2 TX 60 ML Interpretability Data Integration Query Explanations Interaction Debugging Viz Workflow Debugging Application Design Search Replication and Reproducibility Multi-Application Linking Why-not Analytics Interactive Data Profiling Visualization Deconstruction and Restyling Interaction By Example What-if Provisioning Iterative Analytics Interactive Data Cleaning Interactive Query Specification Collaborative Communication Action Recovery Sense-Making Meta-Analysis Debugging Auditing Resource Scheduling Network Diagnostics Interactive Visualizations ⨝ γ A F J O Lineage Graph Recent research has shown that provenance-enabled data systems can be fast enough for interactive responsiveness Deep connections between fine-grained provenance and interaction • Can express core visualization interactions • Extends to general interactive applications Expressing interactions using provenance enables holistic optimization