Скачать с ютуб R-Ladies Freiburg (English) - Tidy Text Analysis: Word frequencies & n-grams в хорошем качестве

R-Ladies Freiburg (English) - Tidy Text Analysis: Word frequencies & n-grams

All materials available on the R-Ladies Freiburg Github repo: https://github.com/rladies/meetup-pre... Event was originally live on Zoom with R-Ladies Freiburg, check out our upcoming events here: https://www.meetup.com/rladies-freibu... Taught by Kyla McConnell (@McConnellKyla) and Julia Müller (@JuliaMuellerFr), PhD students at the University of Freiburg in Linguistics. Text is one of the most prominent and extensive data sources in the digital world. It can be an insight into consumer opinions, social media trends, and literary story arches. However, it can be difficult to analyze without specialized tools and commands. The tidytext package and framework makes it easy to go from full text files to clean data and ultimately, analysis! The workshop will start with the building blocks of text analysis: separating a text into smaller chunks, removing punctuation, and identifying frequent words or phrases. If you’ve joined our previous text analysis meetups before, this will be a great recap; if you haven’t, this will bring you right up to speed! We’ll show you all you need to know so beginners to the topic are very much welcome. The focus for the evening will be on going beyond individual words in isolation to n-grams: units of a set number of words (e.g. bigrams like “I think”, “think that”, “that tidytext”, “tidytext is”, “is great”.). This will allow us to get a sense of context about words based on the company they keep and address typical text analysis problems (e.g. "happy" preceded by "not" being given a positive sentiment score). We’ll top it all off with some truly unique visualizations that show the relationship between words, like bigram network plots with ggraph. Materials will be based on the book Tidy Text Mining by Julia Silge and David Robinson (chapter 4).