NEW: Read latest update on social mention sentiment analysis of President Obama’s Middle East speech
Textual analysis of the rhetoric and of the social media buzz created by President Obama’s speech on the future of the Middle East (see transcript of Middle East speech or full video below) reveals that Twitter users are far more optimistic about the plans presented by the president than traditional media or than blog writers. At the same time, semantic and “trope” analysis of the speech reveals that its goal is to recapitulate what is known, rather than to push the action agenda forward. These trends can be revealed with a set of “text mining” tools, some available online, that can capture the implicit rhetorical charge and the “social emotion” created by public communication as it traverses cyberspace.
Social Mention, a “sentiment analysis” and “social media monitoring” site estimates that for each negative tweet related to the “Obama middle east peace” keyword set there were fifteen positive ones. On the other hand, for each negative mention in a media story of the same keywords, there were six positive nods at president’s speech.
In other word, the ratios were 15:1 compared to 6:1. The ratio for blogs was 4:1. While it is hard to draw actionable conclusions, it is clear that the twittering classes (traditionally better educated and more left-leaning) seem to embrace the optimistic vision of the President, even when compared with the left-of-center bias of traditional establishment media. It is also important to note, that the ratios mentioned above leave out neutral statements, which dominate all three realms of discourse. For example, the ratio for neutral/positive statements for twits is 5/1, for news is 2/1, and for blogs 2.5/1.
At the same time, the discourse was vague and focused on story-telling, rather than on prescribing actionable solutions. A simple textual analysis of the speech performed with Tropes, revealed that it tapped in well known concepts. There were no surprising ideas or actions. The most frequently nouns used in the speech were “people” (48 mentions), “region” (37 mentions), “right” (25 mentions), “change” (16 mentions), “country”, “security”, “Israel” (15 mentions), “state” and “palestinian” (14 mentions). The most frequently verbs were “can”, “will”, and “have” and “be” which refer to verbal structures that state facts. Over 31% of the verbs used in speech were stative (that state a property rather than designate an action: begin, can, have, is etc.) and 42% are factive (they assume the truth of the statement: witness, focus, concentrate, refuse) . Performative verbs, which indicate actions were in a tiny minority (0.3). There are a good number of action-oriented verbs, but these are mostly reflexive, referring to the speaker or to an all inclusive “we” (26.2 %), reflecting Obama’s own feelings or what he assumed to be he audience’s implied self-directed actions. The adjectives were dominated were also dominated by well known tropes, such as “democratic”, “peaceful”, “international” or “young”. A typical sentence, that would encapsulate the spirit of the address by using the words most frequently utilized would be “We can change the region by a state of international and democratic security through the power of the young.” It is no wonder, then, that from a speech that was supposed to offer an overall strategic vision for helping the Middle East to come out of its current predicament, the only idea that captured the public attention was the call for Israel to make peace on the basis of the 1967 border. An idea that in fact is nothing other than a repetition of Saudi Arabia’s last minute attempt in 2002 to save the Palestinian-Israeli peace process by guaranteeing Israel the same 1967 borders.
We will continue monitoring the ripple effects of the speech in social media and will update the numbers offered here with new data as it emerges. A future post will analyze the overtime coverage of the speech, especially when compared with concurrent and competing topics, such as Al Qaeda’s latest pronouncements or Pakistan’s stance in the war on terror.
The figures are consistent with those provided by other social media monitoring sites, such as Opinion Crawl, which indicates a similar greater enthusiasm for president Obama’s speech.
Social mentions culls a large variety of sources, in total 80 media sources. While its sentiment analysis algorithms are not public, they are in tune with those used by other social media monitoring tools (twitterweather, opinoncrawl, twittersentiment, etc.), which look at how polarized statements are in terms of adjective use and sentence structures.
Sentiment analysis has emerged in the last couple years as a hot area of research and business intelligence. Of the many sentiment analysis services available online, the most notable, besides the one already mentioned, are summarized in this spreadsheet created by a group of Stanford students, creators of their own Twittersentiment app. Feel free to add and edit it.