In this work, to classifiedclassify the tweets using Rstudio on Twitter. These tweets are used to analyze positive and negative tweets to make polarity scored. The polarity scores result data could be extracted from Twitter. Extracted data are prone to have security issues. Hacking the data easily and changed the score results lead to lots of issues like affecting the economic status, company brand, and reputation of firms. Salsa20/4 is faster encryption because of the quarter round, which is also better security. The novel algorithm is proposed by modifiedmodifying the Salsa20/4 to enhance further the security of the accumulated data. The proposed algorithm SRB18 has two stages. The first stage of SRB18 is the secret key, and the second stage of SRB18 is the diagonal elements in N-1 steps. SRB18 results show that the encryption algorithm takes high execution time because of the secret key and high data security when compared to Salsa20 variant. The advantage of the SRB18 algorithm is a decryption algorithm, which is a huge time taken when compared to the Salsa20/4 algorithm. In the future, the secret key will be asa fraction value and still need to add more operations of the data security.