Machine Learning is one of the apogees growing data science having tremendously large domain of applications. It is a subfield of data science; having breathtaking approaches in the recent years and will surely grow with a rapid rate. According to current scenario, the use and demand of social media is increasing tremendously. People are relying more on internet, social media and thus the size of the data is becoming huge day by day. As the size of data is extremely huge, so there is a frightening rate to analyze the data and to extricate the useful data various technologies like big data, cloud computing, machine learning, deep learning, data science are coming up in the market. In this paper, we are analyzing tweets. From twitter, we have collected a large number of tweets, tokenization is performed using anaconda, an integrated development environment of python and then training is done on the tokenized tweets to identify the useful data. Spyder is yet another powerful technical environment of python. It provides advanced features of data analysis as well as excellent visualization.