Matisse: A visual analytics system for exploring emotion trends in social media text streams

Dynamically mining textual information streams to gain real-time situational awareness is especially challenging with social media systems where throughput and velocity properties push the limits of a static analytical approach. In this paper, we describe an interactive visual analytics system, called Matisse, that aids with the discovery and investigation of trends in streaming text. Matisse addresses the challenges inherent to text stream mining through the following technical contributions: (1) robust stream data management, (2) automated sentiment/emotion analytics, (3) interactive coordinated visualizations, and (4) a flexible drill-down interaction scheme that accesses multiple levels of detail. In addition to positive/negative sentiment prediction, Matisse provides fine-grained emotion classification based on Valence, Arousal, and Dominance dimensions and a novel machine learning process. Information from the sentiment/emotion analytics are fused with raw data and summary information to feed temporal, geospatial, term frequency, and scatterplot visualizations using a multi-scale, coordinated interaction model. After describing these techniques, we conclude with a practical case study focused on analyzing the Twitter sample stream during the week of the 2013 Boston Marathon bombings. The case study demonstrates the effectiveness of Matisse at providing guided situational awareness of significant trends in social media streams by orchestrating computational power and human cognition.

[1]  Yale Song,et al.  #FluxFlow: Visual Analysis of Anomalous Information Spreading on Social Media , 2014, IEEE Transactions on Visualization and Computer Graphics.

[2]  Cecilia Ovesdotter Alm,et al.  Emotions from Text: Machine Learning for Text-based Emotion Prediction , 2005, HLT.

[3]  Eduard Gröller,et al.  The Event Tunnel: Interactive Visualization of Complex Event Streams for Business Process Pattern Analysis , 2008, 2008 IEEE Pacific Visualization Symposium.

[4]  Yingcai Wu,et al.  EvoRiver: Visual Analysis of Topic Coopetition on Social Media , 2014, IEEE Transactions on Visualization and Computer Graphics.

[5]  Sunghwan Mac Kim,et al.  Evaluation of Unsupervised Emotion Models to Textual Affect Recognition , 2010, HLT-NAACL 2010.

[6]  Ewan Klein,et al.  Natural Language Processing with Python , 2009 .

[7]  William Ribarsky,et al.  LeadLine: Interactive visual analysis of text data through event identification and exploration , 2012, 2012 IEEE Conference on Visual Analytics Science and Technology (VAST).

[8]  Chad A. Steed,et al.  Text Stream Trend Analysis using Multiscale Visual Analytics with Applications to Social Media Systems , 2015 .

[9]  Daniel A. Keim,et al.  Processing online news streams for large-scale semantic analysis , 2010, 2010 IEEE 26th International Conference on Data Engineering Workshops (ICDEW 2010).

[10]  Fangzhao Wu,et al.  OpinionFlow: Visual Analysis of Opinion Diffusion on Social Media , 2014, IEEE Transactions on Visualization and Computer Graphics.

[11]  Stan Szpakowicz,et al.  Using Roget’s Thesaurus for Fine-grained Emotion Recognition , 2008, IJCNLP.

[12]  Daniel A. Keim,et al.  State-of-the-Art Report of Visual Analysis for Event Detection in Text Data Streams , 2014, EuroVis.

[13]  Georgios Paliouras,et al.  ELS: a word-level method for entity-level sentiment analysis , 2011, WIMS '11.

[14]  Reza Zafarani,et al.  Evaluation without ground truth in social media research , 2015, Commun. ACM.

[15]  Lucy T. Nowell,et al.  ThemeRiver: visualizing theme changes over time , 2000, IEEE Symposium on Information Visualization 2000. INFOVIS 2000. Proceedings.

[16]  Carlo Strapparava,et al.  Learning to identify emotions in text , 2008, SAC '08.

[17]  M. Sheelagh T. Carpendale,et al.  A Visual Backchannel for Large-Scale Events , 2010, IEEE Transactions on Visualization and Computer Graphics.

[18]  Xin Tong,et al.  TextFlow: Towards Better Understanding of Evolving Topics in Text , 2011, IEEE Transactions on Visualization and Computer Graphics.

[19]  Vijay V. Raghavan,et al.  On modeling of information retrieval concepts in vector spaces , 1987, TODS.

[20]  Mitsuru Ishizuka,et al.  Recognition of Affect, Judgment, and Appreciation in Text , 2010, COLING.

[21]  Véronique Hoste,et al.  Fine-grained analysis of explicit and implicit sentiment in financial news articles , 2015, Expert Syst. Appl..

[22]  M. Bradley,et al.  Affective Norms for English Words (ANEW): Instruction Manual and Affective Ratings , 1999 .

[23]  Ali R. Hurson,et al.  TF-ICF: A New Term Weighting Scheme for Clustering Dynamic Data Streams , 2006, 2006 5th International Conference on Machine Learning and Applications (ICMLA'06).

[24]  Daniel A. Keim,et al.  EventRiver: Visually Exploring Text Collections with Temporal References , 2012, IEEE Transactions on Visualization and Computer Graphics.