Business Case Analysis Finding Tweets With Criminal Intentions

ESCI-ISTTM

Business Case Analysis Finding Tweets with Criminal Crimina l Intentions Int entions -Vikas R, Askok K, Prem Kumar,Swetha Reddy,Ramya Sravanthi,Veenasri,T,Chaitanya

 The document contains a basic study about an analysis done on the Business case of identifying the tweets with criminal intentions and the approaches followed

Business Case Objetives •

•

•

•

 To help the police force to identify the tweets with criminal intentions  To classify the problem as Classi!cation" #egression or $ptimi%ation  To identify the problems that we thin& will become important in sol'ing this  To design a dashboard that will pro'ide the needed insights

C!assi"ation# (ere the main ob)ecti'e is to !nd out any tweets which are tweeted has the intention of doing a criminal act or not So there are only two scenarios possible in which a tweet pic&ed up randomly might fall in after cross chec&ing with some prede!ned models  *Criminal intention tweet+ or *,ot a Criminal intention tweet+  Therefore the problem falls under the category of *CASSIFICATI$,+

To $denti%y the Prob!ems# As per the # and . done" below /0s were considered to be important while sol'ing this business case

1 /$2ME3 Since there are nearly millions of tweets 4ot of .ata which might be in 5B" MB" TB" 6B78 tweeted on a

daily basis 9e ha'e to ha'e a constant chec& on the /olume of the data : /E$CIT;3 Each minute the rate at which the end users &eep on tweeting increases" therefore we should also ha'e a chec& on the /elocity < /A#IET;3 There will be di=erent &ind of tweets which will be tweeted by the users for eg Te>ts" Images" /ideos" use of (ashtags etc ? /A2E3 #eduction and control of ife and Infrastructure damages of the ,ation @ /IS2AIATI$,3 Classi!cation problem is all about partitioning the space Therefore we will be using Scatter lots in order to partition the space where one partition shows the tweets containing *Criminal intentions+ and other partition shows the tweets containing *,o Criminal intentions+

$nitia! Ste&s to be %o!!owed# 1 .ATA C$ECTI$,3 9e will be collecting all the databases from the police department which has the criminal intention tweets and will try to !nd out the commonly used words in these criminal tweets For eg 2sage of Criminal intention words li&e 6uns" Blast" Shooting" (a'oc" 5ill" Bomb" Suicide bomber For e'ery word used in the criminal tweets we will try to plot a bar chart depicting the number of criminal tweets with respect to that word

1: 1   ;ear 1

D

 ;ear :  ;ear <

? :  ;ear < 

 ;ear : 6uns

Blast

5ill

 ;ear 1 Bombs

: .ATA C$ECTI$,3 9e will be collecting the data 4tweets on a regular basis8 and analy%e the tweet with respect to already rede!ned criminal words and classify them Criminal Tweet F 46uns" Blast" Shoot" (a'oc" Bombs" etc8 If the temporary tweet ta&en at random contains either of the abo'e words li&e 6uns" Blast" Shoot" (a'oc" Bombs" etc then the tweet will be considered as a *Criminal Intention tweet+ else *,ot a Criminal Intention tweet+ < /IS2AIATI$,3 Based on the Scatter lot" we can apply a simple logistic regression and di'ide the space into two partitions Any tweet containing the criminal words li&e 6uns" Bombs" 5ill etc will be classi!ed as *Criminal intention tweets+ else they will be placed in the *,on-Criminal Intention tweets+ partition

 '-Va!ues <@ < :@  ;-/alues

: 1@ 1 @  @

1

1@

:

:@

<

Attributes# • • • • • • •

Age 6ender .emographics #eligion Marital Status Employment rior Criminal based tweets

 The abo'e business case analysis is sol'ed only &eeping in mind to chec& whether a random tweet done is a *Criminal Intention tweet+ or *,ot a Criminal intention tweet+ (owe'er some tweets which contain the criminal words li&e 6uns" Bombs" 5ill etc might not always be Criminal intention tweets since the tweets might be tweeted as )o&es" to show sarcasm and also for fun

(ow to eliminate these &ind of tweets will be done by still applying the deeper Analytics which we are not aware right now