Visual sentiment analysis of customer feedback streams using geo-temporal term associations

Abstract

measures, and maps customer feedback; (2) a novel way of determining term associations that identify attributes, verbs, and adjectives frequently occurring together; (3) a self-organizing term association map and a pixel cellbased sentiment calendar to identify co-occurring and influential opinion; and (4) a new geo-based term association technique providing a key term geo map to enable the user to inspect the statistical significance and the sentiment distribution of individual key terms. We have used and evaluated these techniques and combined them into a well-fitted solution for an effective analysis of large customer feedback streams such as web surveys (from product buyers) and Twitter (e.g., from Kung-Fu Panda movie reviewers).Keywords: Customer Sentiment Visual Analytics, Term Association, Geo-Term Association, Pixel Geo Map, Key Term Geo Map, Pixel Calendar. Introduction MotivationWith the rapid growth of social media, the number of customer comments available to corporations, business owners, and service managers interested in obtaining customer feedback is larger than ever. In addition to the traditional web survey, Twitter is a relatively new phenomenon that has the potential to generate massive amounts of customer comments. However, the language of the tweets is more casual than that of web reviews. Tweets are by definition short (maximum 140 characters) and tend to contain a significant number of abbreviations. The enormous size of the customer feedback data stream, the diversity of the comments, and the uneven distribution of feedback over time make sentiment analysis of this data very challenging.A set of common questions arises in the analysis of customer comments from surveys and other online data streams. Are there aspects of location or geography that impact how a product or service is received by customers? Does a product or service work better for people on the coasts compared to people living in the interior of the country? Does it make a difference if the customer lives in a remote area rather than in a high density urban setting? Is the product or service more appreciated in certain states or cities? What are the important features, attributes, and associated context terms, such as products, timely delivery, channel vendors, product quality, or past experiences that our customers want? How significant terms (content-bearing words, e.g., compound nouns, adjectives, and verbs) are best extracted and presented to business managers so they can understand the results (positive versus negative)? Business managers want to see the sentiment value of the review, but they also want to know the important terms in the context of a review. Furthermore, how to visualize reviews in a dense area without overlap (e.g., Los Angeles, New York), is also a challenge required to be resolved.To meet the above challenges, we propose a pipeline combining feature-based sentiment analysis and geo-term associations to enable store managers to analyze web survey feedback. As an example, Figure c. Figure 1...

A Survey on Visual Analytics for the Spatio-Temporal Exploration of Microblogging Content

“…There is also a desire to do targeted marketing based on real-time streaming tweet analysis (Ribarsky et al 2014) and to detect geotemporal sentiment patterns, trends and influences in customer feedback streams for live alerts (Hao et al 2013). …”

The State of the Art in Sentiment Visualization

“…Hao et al . [HRJ*13] present four novel visualization techniques for customer feedback analysis: pixel sentiment geo map, key term geo map, pixel‐based sentiment calendar, and self‐organizing term association map. Some other examples for this category include map of opinion clusters introduced in the work by Oelke et al .…”

“…Most of such techniques support details on demand, typically by providing the user with a detailed (text) view for the selected timeline [DQP*12, BRT*14] or map [LWW*16, LSB*16] region. Another option is to display specific analysis details for a selected/hovered item [HRJ*13, TSA14, CCYT15].…”

SentiCompass: Interactive visualization for exploring and comparing the sentiments of time-varying twitter data

“…happy, sad, stressed) conveyed by their tweets using text analysis methods. To visually facilitate sentiment and affective analysis, many previous works [15,7,6] abstracted "sentiment" as a unidimensional variable ( i.e. from negative to positive) for visualizations.…”