Capitalizing, qualifying and quantifying the tweets about NSA on Twitter
This analysis is indicative and provides directional guidance on how NSA is perceived across the world.
2. Objective, Timeframe and Source of Data
Objective: Capitalizing, qualifying and quantifying the tweets about NSA on Twitter
Time frame : October 2013 to November 2013
Geography: Tweets from across the world were considered for the data collection
Language: Tweets which are only in English were selected after filtering other national languages
Note :
• For the analysis, open source Social Media monitoring tool Trackur was used.
• This analysis is indicative and provides directional guidance on how NSA is perceived across the
world.
• The data used for analysis solely dependent on the tool used, volume of the data captures and
timeframe considered for the study and human intervention to ensure data has consistency and
credibility to derive inferences
Source: Trackur – http://www.trackur.com/
3. How Noise was excluded from Data
1. Some tweets were in language other than English which were not considered for data analysis
Example: Documentos da NSA so distribudos como
segurana.https://twitter.com/BrendaGwen_/statuses/389833378029395968
2. Few Tweets were repeated multiple times which were directing us to the same source by the same
user hence they were not considered for data analysis
3. Few profiles from the data were either suspended/ page doesn’t exist which are represented below:
A) Example: The National Security Agency is gathering email...
https://twitter.com/pRaTiK_wow/statuses/389944722925817856
B) Example: NSA Discloses Cellphone Location Tracking Tests:..
https://twitter.com/TriumphCISO/statuses/385795119053430786
5. Data Collation and Cleansing
Negative Neutral Positive Total
Initial data from the tool 155 603 24 782
Final data after cleansing 168 489 18 675
155
603
24
168
489
18
Negative Neutral Positive
Data Cleansing
Initial data from the tool Final data after cleansing
782
675
Intial data from the tool Final data after cleansing
Total Tweets
• After reading through the tweets we could infer that there are 107 tweets which can be considered
as null/invalid based on the language, repetitive tweets, inactive links and suspended accounts.
• We have listed all these invalid tweets in our reference sheet with the tab name as “Tweets –107”
6. Sentiments of the Overall Posts
• The overall sentiment of posts were neutral considering the fact that they was more viral in terms of
what and how NSA has been carrying out their operations, the opinions of the public are also captured
along with the categories which are shared in the proceeding slides in detail
Total Number of posts N=675
Negative, 25%
Neutral, 72%
Positive, 3%
168
489
18
Negative Neutral Positive
7. Overall Posts by Category
• Most of the tweets were primarily having mentions about Obama/Govt. which has a direct correlation to
NSA in the tweets that were gathered followed by the discontent by the companies which have the data
of the customers which were accessed by NSA without their approval
Total Number of posts N=675
55
15
21
18
566
0 100 200 300 400 500 600
Companies
Countries
Edward Snowden
NSA Employees
Obama/Govt
8. Qualitative Sentiment Analysis
• Most of the tweets were centered around the way in which NSA/Obama administration has handled the
entire situation, it has received most of its tweets under the sentiment as "Neutral" considering the
articles which were published by the various news corporation for the knowledge of the pubic.
Total Number of posts N=675
Companies Countries Edward Snowden NSA Employees Obama/Govt
8
11
4
7
138
47
4
15
11
412
2
16
Positive
Neutral
Negative
10. Overview #NSA Tweets
• The National Security Agency of the United States (NSA) spied 35 phone calls from world leaders
• The National Security Agency recorded information about more than 124 billion phone calls in one
month
• The National Security Agency has been harvesting millions of e-mail addresses from contact lists of
users of sites such as Facebook, Gmail, Hotmail and Yahoo.
• French Condemn Surveillance by N.S.A.
• The NSA Hacked Former Mexican President Felipe Calderon’s Email While He Was In Office
• The National Security Agency, the controversial U.S. surveillance agency, has had a tough few weeks,
as new allegations from leaked documents have angered more allies.
• Edward Snowden: US would have buried NSA warnings forever
• Snowden claims Russia, China did not get NSA documents
• National Security Agency has had limited success in cracking Tor communications
• Government Shutdown Slows NSA Surveillance Reform Efforts:
• NSA’s vast new Utah data hub suffering from ‘meltdowns’ – report
• White House confirms National Security Agency (NSA) chief to step down
15. NSA — Positive tweets
White House press secretary Jay Carney
confirmed that National Security Agency
Director Gen. Keith Alexander would step down
early next year but downplayed suggestions his
departure had anything to do with recent
damaging leaks from Edward Snowden.
19. Manual intervention - Changed the sentiment from
“Positive” to “Negative/ Neutral” –
Positive to Negative Positive to Neutral
Positive to Negative Positive to Neutral
20. Conclusion
Neutral
Based on the analysis carried out on the data we have derived the following:
• Most of the articles had their reference to the twitter account of the news corporation which
twitter users (individuals) have leveraged for retweeting so the sentiment remains to be neutral
• There are few instances in which the tool has captured the tone in a way which the user never
intended the tweet to be so we had to change the sentiments, the examples for which were
already in the earlier slides
• Few of the twitter users have dared to tweet about NSA directly in a tone which was Negative
and few have taken an alternate route of using Sarcasm but most of them refrained from
commenting hence we have most of Tweets on an Neutral Tone from the data .
In conclusion we infer that the overall tweets are mostly having a “neutral "sentiment towards NSA