Htgotcode/Twitter-ZA-News-Data-Collection-and-Text-Mining

South African news Twitter accounts will have their timeline analyzed on how their sentiment & topics change over time. # Twitter-ZA-News-Data-Collection-and-Text-Mining South African news Twitter accounts will have their timeline analyzed on how the topics change over time. # Requirements: ## Minimum requirements: Identify a set of twitter accounts for media agencies in South Africa, extract their latest 3200 tweets, identify relevant tweets from this set, for those tweets collect relevant interaction data (likes, retweets, and accounts retweeted etc.). With this dataset, provide a descriptive account of 1) how reporting has progressed over time, 2) what topics have been reported on, 3) the tonality/sentiment of this reporting (choosing the most appropriate lexicon in your view), 4) the tonality/sentiment for specific topics, 5) how each of these differ by a) time and b) media agency. ## Additional requirements: In addition to the minimum requirements, to provide further insight, you are expected to identify additional relevant sources of data that might further contribute to our understanding of the twitter conversations in South Africa surrounding the pandemic over the last 6 months and conduct appropriate analyses with the sources. These analyses may be similar to the primary analyses or, optimally, they may go beyond the minimum requirements. For data, you might consider alternative data from twitter that is relevant, or you might consider if there is anotherplatform or set of platforms that provide relevant data (for instance media reporting or other discussion fora). ## Resources for text analysis: - Text Mining with R tidytextmining.com - Text as Data cbail.github.io - sentimentr package: cran.r-project.org - vader package cran.r-project.org - udpipe package: cran.r-project.org - LDA Tuning: cran.r-project.org ## Expectations: This is simultaneously a …