SPIN
Twiter based, anonymous (Non API-dependant), artificial intelligence for mass political profiling and covert persuation.
PROGRAM FLOW:
- Unauthenticated twitter client downloads tweets from target username(s)
- Supervised learning sentiment analysis is performed on the tweets to determine target's personality.
- Tweet text tokenizing, character filtering.
- Penn bank compliant Part Of Speech tag recognition.
- P.O.S based Context-dependant word stemming.
- Supervised topic recognition, tweet categorization.
- Topic specific sentiment analisys to determine target's political orientation on the Nolan chart.
- (Optional) Propaganda function call to influence target based on learnt behaviour.
Key concept:
If wanted to benefit (A) wing over (B) wing, pro BLANK vote propaganda will be sent - in some manner- to those who tend to the opposite (B) wing, and pro VOTE propaganda to those who tend to (A) wing, resulting in a higher proportion of votes for (A) wing.
Example output:
python3.6 deploy.py -t potus -b left -c 100
___ ___ ___
/\ \ /\ \ ___ /\__\
/::\ \ /::\ \ /\ \ /::| |
/:/\ \ \ /:/\:\ \ \:\ \ /:|:| |
_\:\~\ \ \ /::\~\:\ \ /::\__\ /:/|:| |__
/\ \:\ \ \__\ /:/\:\ \:\__\ __/:/\/__/ /:/ |:| /\__\
\:\ \:\ \/__/ \/__\:\/:/ / /\/:/ / \/__|:|/:/ /
\:\ \:\__\ \::/ / \::/__/ |:/:/ /
\:\/:/ / \/__/ \:\__\ |::/ /
\::/ / \/__/ /:/ /
\/__/ \/__/
Red Team Intelligence
https://github.com/zadewg/Election-Meddling
[*] Target: potus (Washington, D.C.) 23763183 followers
[*] Benefit left wing
[*] Collecting ~100 Tweets
.
.
.
Tweets len ID Date Likes RTs SA
751 .@realdonaldtrump: "We're meeting with the Cha... 94 1002633696086593536 2018-06-01 19:31:15 11748 2559 0
752 Congratulations to Vice Admiral Karl Schultz o... 195 1002628738146865154 2018-06-01 19:11:33 5983 1009 1
753 Admiral Paul Zukunft has done an outstanding j... 156 1002628407807619073 2018-06-01 19:10:15 5246 765 1
754 Honored to join @POTUS this morning at the @US... 269 1002627439892271106 2018-06-01 19:06:24 6754 1131 1
755 This is the lowest unemployment rate in more t... 170 1002620046764539904 2018-06-01 18:37:01 6033 1515 1
756 President Trump is working hard to create a th... 239 1002612998433923078 2018-06-01 18:09:01 4917 1226 1
757 The @NYTimes: “We Ran Out of Words to Describe... 148 1002598904498704385 2018-06-01 17:13:00 8885 2674 1
758 Record JOBS DAY!! 223K jobs added. Unemploymen... 280 1002602715820756993 2018-06-01 17:28:09 40487 9077 1
759 President Trump proclaims June 2018 as Great O... 190 1002595130304598016 2018-06-01 16:58:01 5415 1114 1
760 During National Homeownership Month, we affirm... 231 1002588594207383552 2018-06-01 16:32:02 3795 766 1
761 President Trump continues to bolster job creat... 158 1002585315549302784 2018-06-01 16:19:01 4919 1071 1
762 The United States is a nation whose identity, ... 276 1002584816016125954 2018-06-01 16:17:01 4473 941 0
763 “Throughout the Admirals tenure, he poured his... 221 1002580989497860096 2018-06-01 16:01:49 5350 1069 1
764 “We are gathered together today for a truly sp... 254 1002579894692265984 2018-06-01 15:57:28 5672 1086 1
765 “I am thrilled to be here with the brave men a... 246 1002578420306345985 2018-06-01 15:51:37 9402 1805 1
766 President Trump participates in the U.S. Coast... 144 1002565818696327168 2018-06-01 15:01:32 4322 875 1
767 Under President Donald J. Trump, the US unempl... 189 1002563670306672640 2018-06-01 14:53:00 7268 1819 1
768 President Trump Proclaims June 2018 as Nationa... 106 1002550752127135744 2018-06-01 14:01:40 3536 685 0
769 At 3.8%, the unemployment rate has fallen to i... 288 1002558333973393408 2018-06-01 14:31:48 12898 3026 1
770 Jobs! Jobs! Jobs! Once again, the economy is g... 225 1002548108755075074 2018-06-01 13:51:10 17077 3980 1
771 Looking forward to seeing the employment numbe... 70 1002510522032541701 2018-06-01 11:21:48 75725 11371 0
772 Lucia Mutikani in Reuters: "U.S. consumer spen... 192 1002314283663941632 2018-05-31 22:22:01 3210 750 1
773 The U.S. condemns the attacks by pro-governmen... 310 1002312027598217216 2018-05-31 22:13:04 7096 2072 1
Percentage of positive tweets: 60.51282051282051%
Percentage of neutral tweets: 23.974358974358974%
Percentage de negative tweets: 15.512820512820513%
Average subjectivity: 5.0%
Detected 116 mentions.
Hillary sentiment: 0
Trump sentiment: 0.07870129870129869
Cruz sentiment: 0
Bernie sentiment: 0
Obama sentiment: 0
Rpublican sentiment: 0
Democrat sentiment: 0
Guns sentiment: 0
Immigration sentiment: 0.07575732734614314
Employment sentiment: 0.21539018334606566
Inflation sentiment: 0
Minimum Wage Up sentiment: 0
Abortion sentiment: 0
Govenment Spending sentiment: 0
Taxes Up sentiment: 0
Taxes Down sentiment: 0.026713048855906002
Death penalty sentiment: 0
Health care sentiment: 0.17470238095238097
LGBT sentiment: 0
Environment sentiment: 0.06666666666666668
Welfare sentiment: 0.16666666666666669
Social sentimnt: 0
Labor Union sentiment: 0
Drugs sentiment: 0.1463888888888889
War sentiment: 0.1733061354489926
[*] Action: Send Pro-Vote propaganda.
TO DO:
- Code styling, optimization. Propaganda function.
The program currently employs Textblob's default naive bayes clasifier for sentiment analysis.