LDA topic modeling using LDAvis

Topic modeling of user’s privacy associations on Twitter and Instagram

As part of my dissertation, I examined user generated content (UGC = Posts & Tweets) on Twitter and Instagram. More specifically I focussed on users’ privacy associations. I evaluated the UGC with help of two topic models (LDAvis package in R).

A very helpful description of the modeling process can be found here: https://ldavis.cpsievert.me/reviews/reviews.html

After modeling, html-docs were created via pyLDAvis in order to embed the models in this post.

Here I am solely presenting the interactive results of the LDA models for Instagram and Twitter. (Interpretation can be found in my dissertation)

Note: Be ethical about user data.

Requirements

library(lda)
library(LDAvis)
library(tm)
library(dplyr)
library(textclean)
library(stringr)
library(jsonlite)
library(rjson)

LDA topic model Instagram

LDA topic model Twitter