Tags

, , , , , , , , , , , , , ,

I’ve started a Storify project, of uncertain breadth and intention, to document fifty tweets from among Twitter users in Syria, with a loosely geographically stratified sample based on pre-war population estimates (and at that encompassing only up to 43% of Syria’s population).

The first three days can be found here:

http://storify.com/shefaddin/50-tweets-from-syria-august-28-2013

http://storify.com/shefaddin/50-tweets-from-syria-august-29-2013

http://storify.com/shefaddin/50-tweets-from-syria-august-30-2013

My translations range from appalling to acceptable.  In part I do this pretty late when I’m exhausted.  And in larger part my Arabic isn’t brilliant, nor do I have any specific knowledge of or connection to Syrian Arabic.  So at worst the English aspect will improve over time.

In many cases, however, a full translation is not necessary to gauge sentiment.  This brief post is an examination of some extracted data from the first three days:

Sentiment : August 28, 2013

Aug 28 - SentAnalysis

Sentiment : August 29, 2013

Aug 29 - SentAnalysis

This project is too nascent and uncorrelated to draw sweeping conclusions.  However, as discussion of an American strike got underway, it is possible that social media criticism of Bashar became more muted in favor of, generally, less talk about his merit specifically.  However, positive sentiment may also have risen.

Wordcloud from August 28th (probably slightly off, and excluding the keyword ‘Bashar’ and ‘al-Assad’)

Aug28-NoBashar

Wordclouds from August 29:

Wordcloud excluding ‘Bashar’ and ‘al-Assad’

Aug29-NoBashar

Full chart

Aug29

Wordcloud only excluding keyword ‘Bashar’

Aug29-NoBasharJust

Day 3

Aug 30 - SentAnalysis

This trend may have continued into August 30th, in which negative sentiment fell somewhat and positive sentiment remained constant, whereas neutral sentiment rose slightly.

August 30th Wordclouds:

Full Wordcloud

Aug30

Wordcloud excluding keyword ‘Bashar’

Aug30-NoBasharJust

Wordcloud excluding ‘Bashar’ and ‘al-Assad’

Aug30-NoBashar

It is unclear yet to what extent a stable social media signature for these identities exists.  Thus it is unclear to what extent sentiment is connected with external developments.  However, if there is no escalation over the next day or two, it seems likely that negative sentiment will level off along with positive sentiment, which could decline slightly.  Neutral sentiment is likely to continue to rise slightly before itself leveling off.  Galvanization is likely to occur with further action on either side.

3daychange

However, it is not clear to what degree these users are representative of Syrian society more broadly.  I attempted to capture as random a sample as possible, but these are still internet users – all tweeting in Arabic.  It cannot yet reach most of the rural population, presumably; and representativeness is also dependent on infrastructure conditions, just as content may be dependent on political control.

I was personally surprised by the degree of free, open, and unabashed criticism of President al-Assad.

What I will say is that governmental announcements of serious considerations of a military strike against Syria, in particular actions by the American Presidency, have had a dual-pronged effect in social media terms.  On the one hand, there is some evidence that the threat of an attack has produced defections, although of unclear significance and among rank-and-file soldiers.  On the other hand, it may be muting criticism of President al-Assad in favor of the broader anti-neocolonial or anti-interventionist discourse.

I suspect the Administration is aware of this and has taken note, alongside the intelligence community.  It would make any significant ground action more difficult, although none is being considered.  However, it is also certain that any strike will affect the balance of power, although this may suggest that it could actually strengthen Assad’s position, rather than give an inherent boon to rebels as they exploit the situation – as some have suggested.

 

Update: Sept 1:

Sentiment

Advertisements