Today's world has witnessed the information has increased through the social media. These media users tweet on separate topics from anywhere through the net. Tweets area collected, and analyse positive and negative messages, and to drive polarity scores and additionally to predict future trends. These polarity scores might be extracted from Twitter. These information hacked easily, and affect the score results. And make issues, like economic standing, company brands, and therefore the reputations of companies. To tackle these problems, Daniel Bernstein planned to the implementation of Salsa that has quicker encryption because of the district round with better data security. In the current work, a novel methodology is proposed by modifying the Salsa20/4 to increase the security of the polarity scores, which is an important requirement of the current world. The proposed methodology has swapping the co-prime numbers of the column operations. The proposed method has calculate the running time and compared with the existing method; and also provide more security of the analyzed data while to comparing to Salsa20/4 method. © 2020 Elsevier B.V., All rights reserved.