Twitter sentiment analysis on Nigerian Telecommunication Networks.

The task required in here is to know the sentiments( Opinion or emotional artificial intelligence) on Nigerian Telecom providers handle. I will be scraping tweets tweeted by users been directed to their ISPs ( Internet Service providers ) concerning issues they facing about a point in time. I am using the Nigerian network as case study it could work with some other areas too.

Approach : Hybrid Naive Bayes approach ( Combination of Machine learning and Lexical dictionary )

Platform : Twitter

Tool : R programming language ~ Often used by Data science students ( Could be done with Python too)

Twitter Handles :

Airtel - @airtel_care
MTN - @MTN180
9mobile - @9mobilengcare
Glo - @glocare


Other areas sentiment analysis can used :

Product and Service reviews
Reputation Monitoring
Result prediction
Decision making

etc


Well, Enough of Hippie-Dippie side talk - Let get started !


CC: Seun , Lalasticlala

Glo Nigeria: https://imgur.com/a/e0Y46



Picture 1 - Shows RAW data with users Tweets and Opinion :- Problem, Joy, Awkwardness, Negative feedback etc



Picture 2 - Shows the cleaning up of data :- i.e Data Munging! Removing links and unneccessary junks that wouldn't add up to our final result.





Picture 3 - Visualization - After Categorizing every word to it group/class of sentiments we have the report above.



Picture 4 : These are the words that makes up the visualization i.e WordCloud. The bigger it gets more often it gets used.

Thanks all. Based on the report one good deduce the best network when seeking to pick up a network provider for your home use. It will also help the Network provider to help fix problems areas where users are lamenting.

I hope it helps