Telecom network often encounters large number of tweets based on the user experience for a network. This huge amount of raw data can be used for industrial or business purpose by organizing them according to our requirement and processing. The aim of this paper is to address the social media review challenges in telecom companies.  The methods include; extracting tweets, analyzing them and segregating them into various categories to help the company understand the concerns of their customer. This can help save millions and prevent customer churn. In other to build a robust model, the dataset was pre-processed by checking and removing Nan values. After the pre-processing, stage the cleaned data was tokenized. In tokenization process, each word was divided into tokens, for easy training. After the tokenization process, we performed an exploratory data analysis on the dataset to understand the pattern of the dataset. After the explorative data analysis stage, we trained a random forest classifier to predict/classify the customer’s satisfaction into positive, negative, and neural.