The Sentiment Analysis Symposium program for Tuesday, October 30, 2012 is outlined in the agenda that follows. (Also see the pages for the optional October 29 Practical Sentiment Analysis tutorial and Introduction to Social Media Listening class.)
|Morning Session -- Tuesday, October 30|
|8:30 am-8:40 am||
Chair's Welcome: The Sentiment Spectrum
Seth Grimes, Alta Plana Corporation
|8:40 am-9:20 am||
Keynote: Sentiment Driven Behaviors; Sentiment Driven Decisions(mouseover here for description)
Executives are eager to use sentiment as a deciding factor in a wide spectrum of business decisions-- from investing further marketing dollars to developing new products to enhancing call center operations. The origin of this idea is well-founded in the relationship between market sentiment and price development in the world of investing. Sentiment analysis in social media has made substantial headway in applying a modicum of structure to a massive, unwieldy, unstructured set of blogs, forums, microblogs, and social networks. However, the precise relationship between sentiment and subsequent (market) behaviors is variable. Structure does not necessarily beget behavior, and thus proves precarious for business decisions. This talk will review the outcomes of specific business decisions based on sentiment and address three major factors as to why the relationship is as of yet undefined.
Kate Niederhoffer, Knowable Research
|9:20 am-9:45 am||
Up Close and Personal: Social Media Insights and the Mind of the Consumer(mouseover here for description)
Augmenting a structured quantitative survey with social media data from your respondents can strengthen your typical survey analysis with rich qualitative insight. As part of your survey data collection effort, you also bring in your respondents' social media data in real-time. Combining your respondents' social media interactions with their structured This technique will show you how to establish in-depth, sophisticated segmentation based on social media narratives gathered at the same time you do online survey research.
Carol Haney, Toluna
|9:45 am-10:10 am||
Emotions Affect Markets in Predictable Ways(mouseover here for description)
Where there are markets, there are emotions. Where there are emotions, there are cycles. Where you understand cycles, you profit.With modern economic and financial theory increasingly looking towards behavioural and crowd-based models to explain cyclical booms and busts, the ability to quantitatively assess broad market sentiment can be a valuable indicator to help guide successful trading and investment strategies. The world of big data is a new market data frontier that can provide trading firms with a real edge by helping to model behavioural economics. News and Social Media Sentiment toolsallow financial professionals to analyze news and social media in real-time to convert the volume and variety of professional news and the internet into manageable information flows that drive sharper decisions.
Aleksander Sobczyk, Thomson Reuters
|10:10 am-10:20 am|
|10:20 am-10:40 am||Break|
|10:40 am-11:10 am|
|11:10 am-11:30 am||
Using Sentiment Analysis to Make Net Promoter More Actionable(mouseover here for description)
J.D. Power uses sentiment analysis of survey verbatims combined with the Net Promoter Score (NPS) to better understand the customer service experience for a major US bank and provide actionable recommendations for improvement of the branch experience. In this talk, using this real-world example, we will explore the relationship of sentiment to NPS and discuss how to use them together to uncover opportunities for improvement that would have otherwise been missed.
Bill Tuohig, J.D. Power and Associates
|11:30 am-11:50 am||
Amplify Sentiment by Measuring Impact(mouseover here for description)
Dow Jones' Insight Media Index is a methodology for weighting sentiment by message strength, audience, placement, and other criteria chosen by the user. By choosing up to a dozen weights, a client can analyze large volumes of coverage to distinguish between kerfuffles and perfect storms. I'll show how the methodology works, discuss its applications and limitations, and present some real-world case studies.
Barry Parr, Dow Jones
|11:50 am-11:57 am||
Using Social Sentiment to Track Real Time Public Opinion
Rob Bailey, CEO, DataSift
|11:57 am-12:25 pm||
Analyzing Weibo, the Chinese Twitter (mouseover here for description)
I will share my experiences in analyzing Weibo posts. First, I will cover the challenges of working with Chinese text. Then, will share some of my personal solutions as well as available resources. Finally, I will point of some Weibo-specific shortcuts that can be utilized.
Ken Hu, Soshio
Dissipating Ambiguity: Direct Extraction of Sentiment from Social Networks (mouseover here for description)
Modern systems of social network sentiment extraction lean heavily on artificial intelligence scraping and inferring sentiment from public feeds and streams. There is an alternative. Startup Swipp, Inc. has a new approach to sentiment collection and analysis that eliminates ambiguity, adds additional layers of value, and actually spurs the community to disclose more of what they think and feel -- all while making the broader user and commercial networks smarter. This new angle will not only allow sentiment to surface for more direct analysis, but will actually motivate and inspire new behavior, shaping it along the way.
Dr. Erin Olivo, SmogFarm
Analysis and Visualization of Social Media Content (mouseover here for description)
This presentation briefly summarizes value of social media content analysis and importance of data visualization. Then, it describes individual entities used for specific analysis and states some of tools that are being used. The main centre of the presentation is dedicated to different possible ways of visualisation of web content discussions (with focus on social media discussion); it focuses in particular on ways to show through visualisation relationships between behavior and number of users, topic discussion and sentiment over time. New ways of graphical illustrations, e. g. time axis in circles or different types of polygons, will be presented. The development of analytical and visualization portlets will be also mentioned. Outputs from existing content will be presented as well.
Tereza Pařilová, Masaryk University
Beyond Sentiment: The Next Generation of Social Intelligence (mouseover here for description)
As brands continue strive to understand their sentiment in social media, a new generation of social intelligence is rapidly emerging. Emotions, intention, message types and other new classifiers are providing a fuller spectrum of insights and meaning that can be game-changing. This session will discuss, briefly, some of the new classifiers that are bringing new dimensions to social intelligence.
Mike Moran, Converseon
Mining Big Voice of the Customer Data
Daniel Ziv, Verint
|12:25 pm-12:35 pm||
Building Sentiment Analysis on the Right Social Data(mouseover here for description)
Do you analyze sentiment on social data? What datasources do you use? Where do you get them? The answers to these questionsare just as important as the analysis you run. In this session we will explore the different types of social data available and best practices for consuming such data. After attending this session, you will have an in-depth understanding of the different types of conversations available in different social data sources. We?ll also cover the role that reliability, sustainability and completeness play in yourunderlying data so you can be sure you?re building your analysis on a solidfoundation.
Chris Moody, President & COO, Gnip
|1:30 pm-1:50 pm||
Harvesting Mobile Micro-Slang(mouseover here for description)
Marketers are tasked with understanding sentiment need to understand terminology and language used to describe their product. Approaches in text mining and term analysis that are used in taxonomy and vocabulary development can be used to solve language problems in sentiment analysis. In this session, Jeannine Bartlett provides an example of how to use a text mining tool to beef up a sentiment analysis taxonomy, thesaurus and signal detection strength for mobile micro-slang target audiences.
Jeannine Bartlett, Earley & Associates
|1:50 pm-2:10 pm||
Sentiment Analysis and the Consumer Genome(mouseover here for description)
At Infosys, we've created what we call the Consumer Genome and it could well be the next leap in thinking. Just like its human counterpart, the basic premise of a consumer genome is that certain intrinsic attributes strongly determine consumer behavior. The list includes demographics, connections, influences, interests, needs and buying behavior. That's a marked departure from current consumer understanding practices, with an over reliance on demographic metrics as predictors of behavior. The consumer genome has its own complex DNA, made up of rich information about an individual that can provide unique and compelling insights into that person's consumption behavior. By decoding it, consumer product companies, retailers, indeed all types of marketers can personalize their offerings, channels and campaigns to each and every individual. Call it 'extreme relevance.' The presentation will describe the concept of the consumer genome and will explain how semantic analysis and artificial intelligence on a big data platform are leveraged to realize its promise.
Vaidyanatha Siva, Infosys Ltd.
|2:10 pm-2:30 pm||
A Tailor-Made One-Size-Fits-All Approach to Sentiment Analysis
Diana Maynard, University of Sheffield
|2:30 pm-2:50 pm||
Assess and React to Market Situations Using Social Data(mouseover here for description)
This session will show how a variety of data sources, including revenue, brand tracker, and social media work together to evaluate brand reputation and support recommendations for organizational reactions. Specifically, examples of sentiment analysis and buzz volume are used to show the impact and context that social data provides.
Liz Keck, American Cancer Society
|2:50 pm-3:10 pm||
Cisco's Integrated Sentiment Analysis(mouseover here for description)
Cisco has evolved from analyzing purely traditional media, to developing a truly integrated global approach that now also incorporates social media, industry and financial analyst data, employee conversation and customer verbatims. By minimizing siloed analysis, we are able to understand patterns and drive a cohesive strategy that is relevant to all audiences. In this presentation I will share examples of this integrated analysis as well as discuss some of the challenges we face working with diverse data sets in multiple languages.
Elizabeth Rector, Cisco
|3:10 pm-3:40 pm||Break|
|3:40 pm-4:00 pm||
Google's Text Analytics War on Spam
Mike Moran, Converseon
|4:00 pm-4:40 pm||
Campaign 2012: The Voice of the Voter
Participants present their analyses of candidates and issues, one week before the November 6, 2012 presidential election. Who's winning the online/social sentiment race and Why? What tools and methods can help us hear the voice of the voter? Our speakers give us their takes.
Presentation cancelled due to Hurricane Sandy:
Fear and Loathing on the Social Campaign Trail (mouseover here for description)
What are voters afraid of on the eve of the 2012 election? Fear is one of the most freely expressed forms of sentiment in social media. This "Voice of the Voter" presentation looks social data collected in the final week of October and speaks to the nature and salience of fear among the electorate. Bridging political and computational science, Dr. Shulman will present a frightening array of scenarios predicted in the Tweets and Facebook updates as the final phase of the campaign transpires.
Stuart Shulman, DiscoverText
Twitter, Politicians, and the Voice of the Voter (mouseover here for description)
Political discourse is challenging from a sentiment analysis point of view because political issues are highly dynamic and political language may contain neologisms that do not occur frequently in general purpose lexical sentiment models. Also, the presence of humor, sarcasm, and comparatives may introduce errors in sentiment analysis. In Twitter, these issues are amplified by the use of Twitter-specific features and constrained message lengths. In this session, we will present a collaborative project between the University of Southern California (USC) Signal Analysis and Interpretation Laboratory, USC Annenberg Innovation Laboratory, and IBM Corporation. Our system is relies on manual curation of keywords and hashtags, crowd-sourced annotation, statistical machine learned sentiment models, and a real-time visualization that is ideal for display during live events. We describe our corpus and several experiments using different settings of our sentiment models.
Getting Real(-time) with Live Polling (mouseover here for description)
Which candidate responses in a presidential debate did the audience perceive as dodging the question? What specific moments in a new comedy did TV viewers find hilarious, moving, boring, or painful? Questions like these are hard to answer. Traditional polling provides interpretable data, but it can't give you second-by-second information about people's reactions, and it doesn't scale. Response dials allow fine-grained temporal feedback, but on just one dimension. Social media analysis gets nearer to real time, and on a massive scale, but it doesn't allow researchers to pose specific questions to the audience.
Rishab Ghosh, Topsy Labs
|4:40 pm-5:30 pm||
Presentations and Panel --
Social Intelligence at the Social Centers: eBay, Twitter & Zynga
Twitter, eBay, and Zynga represent three social centers. Twitter invented and epitomizes high-velocity, high-volume, high-value social communications. eBay defined and defines social commerce. Zynga is the social gaming company. Representatives of the three companies will present and then participate in joint Q&A, moderated by social-media strategist and consultant Dr. Natalie Petouhoff.
Competitive Advantage(mouseover here for description)
Your customers are on social media. As you plan sentiment analysis and text mining, remember, so are your competition's customers or your prospects. This talk will cover how you can understand the overall market landscape by social data mining for yours and competition conversations. You can go beyond general competition data to understand difference in customer preference, perceptions among different market players. Social Data for Competitive Analysis can help you setup realistic benchmark numbers to make sense of your internal customer data.
Sudha Jamthe, eBay
Sentiment Analysis for Brands on Twitter(mouseover here for description)
Brands use sentiment analysis on Twitter to listen to and engage with their customers and prospects every day. This talk will share case studies of how some of the more successful brands use sentiment analysis to measure their performance on Twitter; make decisions about media outside of Twitter; and get critical market input for product decisions. In the talk and Q&A afterward, I will also share some of the challenges that need to be addressed in order for sentiment analysis to get to the next level.
Ameet Ranadive, Twitter
Using Player Sentiment to Build Great Games(mouseover here for description)
Zynga games are being played by over 300M people every month across multiple platforms/devices. See how Zynga is using player sentiment as a constant feedback stream into making games players love.
Chris Jones, Zynga