Main content



Loading wiki pages...

Wiki Version:
# Welcome! This is the OSF page Matthew Jordan's Master's thesis in Experimental Psychology at Oxford University, titled "Mean Tweets: The Role of Emotion in the Spread of Moral Ideas on Twitter." --- This project is a replication of [(Brady et al., 2017)][1], a paper out of NYU that demonstrated that the presence of moral-emotional words in tweets makes them more likely to be retweeted. <br> <br> In the spirit of open science and reproducilibty, I have made all my code and data available on this page. The main data set is [filtered_all.RDS][4], located in the **Cleaned Files** folder. You can view these data with [RStudio][2] by entering the following: <br><br> ``` setwd("/Downloads/Tweets") # Wherever you put the file filtered_all <- readRDS("filtered_all.RDS") ``` <br> You can now explore this big data set and re-run the analyses I performed. All of the scripts I used to analyze the data are located in the **R Code** folder. ## Collecting Your Own Data You can collect your own tweets from Twitter's API and run the same analyses to examine moral/emotional language. The steps are as follows: <br><br> 1. Follow [these instructions][3] to create a Twitter App, which you'll need to use `rtweet` and gather data. 2. Modify `Matthew_Thesis_Data_Collection.R` with your personal Twitter App information. 3. Specify keywords and an amount of time to collect for. <br> And you're off! The `rtweet` package saves its output as `.JSON` files. You can now use [Matthew_Thesis_Data_Preparation.R][5] to parse these files and turn them into `.csv` files. <br> The `.csv` files can in turned be read into R and analyzed using the functions in [Matthew_Thesis_Data_Analysis.R][6]. Before you know it, you'll have pretty much completed a Master's thesis. <br> Note that in order to use these two functions, you will also have to download the "Helper" files, which contain useful functions. [1]: [2]: [3]: [4]: [5]: [6]: