Iran Online
Team Members
Richard, Sabine, Orsi, Cameran, Bronwen, Leva, Esther, Demet, Niaz, Jean-Christophe, Erik
Introduction
Research questions
Has repression suppressed voice?
Has the health of the Iranian Web (its content and its users) worsened? (with metrics like: hacked sites, parked sites, deserted blogs, spammy neighborhoods, link farms)
Are we able to diagnose changes in the degrees or levels of expression online?
(are sites blocked, down, etc and do they [keep using] specific [smelly] language?)
Methodology
To study the effect of oppression on expression:
1a) Demarcate Iranian web
- Google ad planner (hit economy)
- Alexa (surfer's paths)
- Balatarin/Donbaleh/Sabzlink (crowd-sourced)
- Likekhor (like economy)
- Google (link economy) (site:.ir & regions search (.coms etc)
- Crawling (cybergeographic imagination)
1b) Demarcate significant 'Iranian' spheres
- Blogosphere - Top bloggers
- Twittersphere - Follow #
- Facebook - groups/pages
- Cloob.ir
- Google Reader
- Friendfeed
2) Characterization of each of the Iranian webs
How to characterize them? What are the features? Inside/outside, farsi, non-farsi, etc. How it ranks, what type of sources, location, language
a) Short history
b) 'How' it works
c) Why it 'works' for Iranians
d) Characterization of (national) web it outputs
laguage it outputs
site locations
audience
*reference your accounts
3) Run all the URLs through the language detection tool.
4) List building techniques:
- concatenate/snowball
- triangulate
- network
Project A) what kind of webs do the platforms/services present?
- Triangulate the lists of URLs (for English and Persian); (Esther)
- Describe the nature of the lists, zoomin in on the URLs shared with other lists, and the kind of websites unique to this list; (Leva, Bronwen, Niaz)
Categories: news (inside or outside iran), government, blogs, etc.
Characterizations of views:
- Ultra-conservative
- Hardliner ('Martyrs of the Supreme Leader')
- Pro-government
- Pro-Khomeini but anti-Ahmedinejad
- Monarchist
- Communist
- Reformist
- Green movement
- Leftist
- Muja (MKO)
- Azeri
- Separatist
- Secularist
- Federalist
- looking at languages and TLDs (Liliana)
Project B)
health of the web (blocked or not, broken or not, parked or not) (Cameran, JC)
Visualization
Treemap
- Filter sets
Data:
- all
- Likekhor
- Alexa
- Doubleclick Ad Planner by Google
- Balatarin
- Donbaleh
Option to add data set?
Layers:
- TLD
- Language
- Site type (characterizations)
Layer 2:
- Blocked
- Broken (does it resolve)
Characterization of Iranian webs
Google ad planner
Alexa
Balatarin/Donbaleh/Sabzlink
Likekhor
Google
Gooder: Localizing Google Reader by Iranian users
Findings
Further Research