Interactive Visualizations for Linguistic Analysis

Transcription

1 Interactive Visualizations for Linguistic Analysis Verena Lyding and Henrik Dittmann Institute for Specialised Communication and Multilingualism, EURAC, Bozen-Bolzano

2 Information visualization The use of computer-supported, interactive, visual representations of abstract data to amplify cognition. (Card et al., 1999) Aim: providing a cognitive aid for the illustration of data and their structure/organization the analysis and manipulation of data Good visualizations use graphics to organize information, highlight important information, allow for visual comparisons, and reveal patterns, trends, and outliers in the data. (Hearst, 2009) 2

3 Some examples Classical ways of visualizing data include: graphs, networks, charts, diagrams, maps, but also text. Text is visual. Text is visual. Text is visual. Text is visual. Text is visual. Text is visual. Text is visual. Text is visual. Text is visual. Text is visual. Text is visual. Text is visual. Text is visual. Text is visual. Text is visual. Text is visual. Text is visual. Text is visual. Text is visual. Text is visual. Text is visual. Text is visual. Text is visual. Text is visual. Text is visual. Text is visual. Text is visual. Text is visual. Text is visual. Text is visual. 3

4 More examples Recently, some visualizations of language data have been introduced, including word clouds, concept galaxies, sparklines, etc. taken from: 4

5 How do visualizations work? Information is transformed into graphics, using visual variables : Taken from Carpendale (2003). 5

6 Constructing meaningful visualizations Follow visualization principles. Sameness of a visual element implies sameness of what the visual element represents. (Tufte, 2006) Follow Gestalt psychology principles of perception, like proximity and similarity. Clutter and confusion are failures of design, not attributes of information. (Tufte, 1999) Don t hide information without indicating what is left out. Present information in context. Overview first, zoom and filter, then details-on-demand (Shneiderman, 1996) 6

7 LInfoVis Linguistic Information Visualization The application of information visualization principles to display any kind of information concerning language and its use. LInfoVis is a specialization of InfoVis. And a particular challenge due to: structure and complexity of linguistic data textual elements The categorical nature of text, and its very high dimensionality, make it very challenging to display graphically. (Hearst, 2009) 7

8 LInfoVis at EURAC Motivation: Development and implementation of visualizations for language data. With focus on the representation and analysis of language resources, in particular corpora. Project running since the end of 2008, by the language technologies group at EURAC; initiated by Chris Culy (now University of Tübingen). By now: Development of several visualization prototypes, investigation of application contexts, visualization tools for linguistic projects. We are generally interested in collaborations! 8

9 Visualizing language data We can distinguish between visualizations for a) the presentation of data b) the analysis of data which can be targeted to 1) language data in context, e.g. KWIC 2) information derived from language data, e.g. frequency lists 9

10 LInfoVis for data presentation Data displays that visually highlight relevant aspects of the data. Some examples for: text collocations occurrences of words over time 10

11 Visualization indicating frequencies of words in their textual context; frequencies are encoded by character size of words. 11

12 Graphic from Wikimedia, by Kai Zimmer, Graph visualization of the word Ziel and its collocations as calculated based on the DWDS core corpus. 12

13 Chart displaying frequencies of the indicated words in newspaper text over a crucial period of time. 13

14 LInfoVis for analysis Visualizations that highlight data characteristics and allow for the interaction with the display to explore the data. Techniques from information visualization: search and filter, for focus and context overview, zoom-in, details on demand multiple views, and brushing and linking 14

15 Interactive visualizations Corpus Clouds (Culy/Lyding, 2009) visualization of corpus query results multiple panels for different types of information interactive features for data exploration 15

16 Interactive visualizations Double Tree (Culy/Lyding, 2010) Double Tree shows a concordance in a compressed form that allows for interactive exploration. 16

17 Interactive visualizations xldds Extended Linguistic Dependency Diagrams (Culy/Lyding/Dittmann, 2011b) xldds support the analysis of dependency structures by providing a set of visual features (such as color, size and shape) for the presentation of relations user controls for focusing on specific information 17

18 Visualizing derived information Linguistic analyses are not only concerned with the linguistic data itself, but also with information about this data. e.g. quantitative analysis Parallel Coordinates (Inselberg, 2009) is a common visualization for high-dimensional data. For linguistic analyses, we have developed Structured Parallel Coordinates (SPC), (Culy/Lyding/Dittmann, 2011a): inherent ordering of the axes advanced methods for filtering, selection and highlighting 18

19 Parallel Coordinates Visualization originally invented by d'ocagne (1885). In modern Information Visualization developed and popularised by Inselberg in 1959 (cf. Inselberg, 2009). Taken from Heer et al. (2010). 19

20 Interactive visualizations SPC - Structured Parallel Coordinates (Culy/Lyding/Dittmann, 2011a) n-grams and frequencies application 20

21 Interactive visualizations SPC corpus comparisons evolution of verb constructions over time and register 21

22 Interactive visualizations SPC ranking comparisons Live demo: using ranking comparisons for sub-corpus analysis automatic re-ordering of the axes by similarity 22

23 Summing it up Information Visualization has a lot to offer for the description and exploration of complex data. Our LInfoVis work aims at combining insights and methods from InfoVis and linguistic research settings. Visualizations need to be put to the test in different areas of linguistic analysis. The users needs and experiences are an important base for: improving existing software and pursuing new directions in LInfoVis 23

24 Thank you! 24

25 Bibliography Bertin, J. (1982): Graphische Darstellungen. Graphische Verarbeitung von Informationen. Berlin/New York: de Gruyter. Card, S. K. / Mackinlay, J. D. / Shneiderman, B. (1999): Information Visualization: Using Vision to Think. San Francisco: Morgan Kaufmann Publishers. Carpendale, M. (2003): Considering visual variables as a basis for information visualisation. Dept. of Computer Science, University of Calgary, Canada, Tech. Rep Collins, C. / Penn, G. / Carpendale, S. (2008): Interactive visualization for computational linguistics. ACL-08: HLT Tutorials. Access date: February 9, Culy, C. / Lyding, V. (2009): Corpus Clouds - facilitating text analysis by means of visualizations. In: Proc. of the 4th Language & Technology Conference (LTC 09), Poznan, Poland, Culy, C. / Lyding, V. (2010): Double Tree: An Advanced KWIC Visualization for Expert Users. In: Proc. of the 14th International Conference on Information Visualization (IV 2010), London, United Kingdom, Culy, C. / Lyding, V. / Dittmann, H. (2011a): Structured Parallel Coordinates: a visualization for analyzing structured language data, In: Proc. of the 3rd International Conference on Corpus Linguistics (CILC 2011), Valencia, Spain. 25

26 Culy, C. / Lyding, V. / Dittmann, H. (2011): Visualizing Dependency Structures. In: Proc. of the Conference of the German Society for Computational Linguistics and Language Technology (GSCL) 2011, Hamburg, Germany, Hearst, M. (2009): Search User Interfaces. Cambridge: Cambridge University Press. Heer, J. / Bostock, M. / Ogievetsky, V. (2010): A Tour through the Visualization Zoo. ACM Queue 8(5). Inselberg, A. (2009): Parallel Coordinates: VISUAL Multidimensional Geometry and its Applications. New York: Springer. d Ocagne, M. (1885): Coordonnées Parallèles et Axiales: Méthode de transformation géométrique et procédé nouveau de calcul graphique déduits de la considération des coordonnées parallèlles. Paris: Gauthier-Villars. Richter, M. (2005): Analysis and visualization for daily newspaper corpora. In: Proc. of Recent Advances in Natural Language Processing (RANLP 2005), Borovets, Bulgaria, Shneiderman, B. (1996): The eyes have it: A task by data type taxonomy for information visualizations, In: Proc. of the IEEE Symposium on Visual Languages, Tufte, E. (1999): Envisioning Information. Cheshire, Connecticut: Graphics Press LLC. Tufte, E. (2006): Beautiful Evidence. Cheshire, Connecticut: Graphics Press LLC. 26

27 Further links and references DWDS Wortprofil 2010 for the word Sprache, Access date: February 9, IN-SPIRE TM visual analysis tools, Access date: February 9, jquery Sparklines by Gareth Watts for Splunk Inc., Access date: February 9, Graph visualization of collocations for the word Ziel by Kai Zimmer, DWDS, Access date: February 9, Tutorial on the Visualization of Linguistic Information by Culy, C. / Lyding, V., presented at the 37. Österreichische Linguistiktagung in Salzburg, Austria, December 7, Visualization as Part of the Linguistic Processing Pipeline by Culy, C. / Lyding, V., presented at the Linguistic Processing Pipelines workshop at the GSCL conference, 29 September 2009 in Potsdam, Germany. Word Clouds by Chris Culy, EURAC, Access date: February 9,