# chen2018sequence

# Sequence Synopsis: Optimize Visual Summary of Temporal Event Data

Concept

Chen et al. develop Sequence Synopsis, a visual analytics framework that enables interactive data exploration of event sequence data. Real world datasets such as EHRs, vehicle fault logs and software error logs, often come with noise such as missing values or irregular values. The proposed framework which is highly tolerant to noise, aims to provide concise yet comprehensive overviews of event sequence datasets, with reduced visual cutter to improve efficiency in event sequence analyses such as patient pathway analysis and predictive vehicle fault analysis. [1]

Implementation

Figure 1 shows an overview of Sequence Synopsis. Section A consists of sequence clusters summarized by the framework, clicking on each cluster will bring up the corresponding summary view. Section B shows a list of events in the selected cluster. Section C shows a circular visualization of event co-occurrence with a focus event as the center point. Section D shows options that allow users to set filtering conditions.

Chen et al. adopt a two-part representation, a generic method to summarize event sequences, the steps are shown in Figure 2.

In this method, patterns are extracted from the original data based on the sequence of events. Corrections are made by inserting or deleting events in order to reduce noise. The authors acknowledge that for this approach there is always a trade-off between reducing noise and balancing visualization readability and information completeness, they follow the minimum description length principle (MDL) to find an optimal point for pattern summarization. The MDL states that "the best model for a dataset results in a minimized description length of it"[2], in this case each pattern is effectively a model. Using the MDL, Chen et al. are able to avoid both overfitting and underfitting models while maintaining the information loss to a minimum.

Related Work

Chen et al. review existing implementations in event sequence visualization. A common approach is to place events along a horizontal time axis, such as Lifelines[3], CloudLines[4] and TimeSlice[5]. Despite revealing detailed information of each event, identifying temporal patterns in multiple event sequences can be challenging.

Frameworks proposed to tackle the aforementioned problem are reviewed in the paper.

LifeLines2[6] provides visualizations at different temporal granularities. EventFlow[7], Trail Explorer[8] and CoreFlow[9] visualize events in tree-like branching structures.

Outflow [10], CareFlow[11] and DecisionFlow [12] condense event sequences into transition graphs and use nodes to depict events at different stages, edges are used to link events in sequence.

(s|qu)eries[13], COQUITO[14] and DecisionFlow [12:1] provide an interface for interactive querying, where users are able to focus on a subset returned from their queries.

Chen et al. further review event sequence visualization systems with data mining techniques such as FP-Viz[15], Frequence[16] and ActiveTree[17].

Data Characteristics

Software logs dataset -- Agavue

  • Data source: An IEEE VIS 2016 Workshop: The Event Event: Temporal & Sequential Event Analysis, http://eventevent.github.io/
  • Size: 58,581 sessions, 2,139,847 events. However, the authors only used 2,221 sessions with an undisclosed number of events.
  • Spatial dimensionality: 2D
  • Temporal dimensionality: static
  • Type: multivariate, 6 variables

Vehicle fault sequences dataset

  • Data source: Experts in vehicle data analytics
  • Size: Logs generated by 261 vehicles in one year, an undisclosed number of events.
  • Spatial dimensionality: 2D
  • Temporal dimensionality: static
  • Type: multivariate

Visualization Techniques

  • glyph

Papers Cited

60

Years Spanned

1996-2017

Application Domain

  • Event Sequence Analysis
  • Event Sequence Visualization

Reference


  1. Chen, Y., Xu, P., & Ren, L. (2018). Sequence Synopsis: Optimize Visual Summary of Temporal Event Data. IEEE Transactions on Visualization and Computer Graphics, 24(1), 45–55. https://doi.org/10.1109/TVCG.2017.2745083 ↩︎

  2. P. D. Gr¨unwald. The minimum description length principle. MIT press, 2007. ↩︎

  3. Plaisant C, Milash B, Rose A, et al. LifeLines: visualizing personal histories. SIGCHI Conference on Human Factors in Computing Systems Proceedings; 1996:221–227. ↩︎

  4. Krstajić, M., Bertini, E., & Keim, D. A. (2011). Cloudlines: Compact display of event episodes in multiple time-series. IEEE Transactions on Visualization and Computer Graphics. https://doi.org/10.1109/TVCG.2011.179 ↩︎

  5. Zhao, J., Collins, C., Chevalier, F., & Balakrishnan, R. (2013). Interactive exploration of implicit and explicit relations in faceted datasets. IEEE Transactions on Visualization and Computer Graphics. https://doi.org/10.1109/TVCG.2013.167 ↩︎

  6. Plaisant C, Mushlin R, Snyder A, et al. LifeLines: using visualization to enhance navigation and analysis of patient records. AMIA Symposium Proceedings; 1998:76–80. ↩︎

  7. M. Monroe, R. Lan, H. Lee, C. Plaisant, and B. Shneiderman. Temporal event sequence simplification. IEEE transactions on visualization and computer graphics, 19(12):2227–2236, 2013. ↩︎

  8. Shen, Z., & Sundaresan, N. (2010). Trail explorer: Understanding user experience in webpage flows. IEEE VisWeek Discovery Exhibition, pp. 7–8, 2010. ↩︎

  9. Liu, Z., Kerr, B., Dontcheva, M., Grover, J., Hoffman, M., & Wilson, A. (2017). CoreFlow: Extracting and Visualizing Branching Patterns from Event Sequences. Computer Graphics Forum. https://doi.org/10.1111/cgf.13208 ↩︎

  10. Wongsuphasawat, K., & Gotz, D. (2012). Exploring flow, factors, and outcomes of temporal event sequences with the outflow visualization. IEEE Transactions on Visualization and Computer Graphics. https://doi.org/10.1109/TVCG.2012.225 ↩︎

  11. A. Perer and D. Gotz. Visualizations to support patient-clinician communication of care. In ACM CHI Workshop on Patient-Clinician Communication. Paris, France, 2013. ↩︎

  12. Gotz, D., & Stavropoulos, H. (2014). DecisionFlow: Visual analytics for high-dimensional temporal event sequence data. IEEE Transactions on Visualization and Computer Graphics. https://doi.org/10.1109/TVCG.2014.2346682 ↩︎ ↩︎

  13. Zgraggen, E., Drucker, S. M., Fisher, D., & DeLine, R. (2015). (s|qu)eries: Visual Regular Expressions for Querying and Exploring Event Sequences. Proceedings of the 33rd Annual ACM Conference on Human Factors in Computing Systems - CHI ’15. https://doi.org/10.1145/2702123.2702262 ↩︎

  14. Krause, J., Perer, A., & Stavropoulos, H. (2016). Supporting Iterative Cohort Construction with Visual Temporal Queries. IEEE Transactions on Visualization and Computer Graphics. https://doi.org/10.1109/TVCG.2015.2467622 ↩︎

  15. Stasko, J., & Zhang, E. (2002). Focus+context display and navigation techniques for enhancing radial, space-filling hierarchy visualizations. https://doi.org/10.1109/infvis.2000.885091 ↩︎

  16. Perer, A., & Wang, F. (2014). Frequence: Interactive Mining and Visualization of Temporal Frequent Event Sequences. Proceedings of the 19th International Conference on Intelligent User Interfaces. https://doi.org/10.1145/2557500.2557508 ↩︎

  17. Vrotsou, K., Johansson, J., & Cooper, M. (2009). ActiviTree: Interactive visual exploration of sequences in event-based data using graph similarity. IEEE Transactions on Visualization and Computer Graphics. https://doi.org/10.1109/TVCG.2009.117 ↩︎

🔄 Last Updated: 7/8/2020, 9:48:18 PM