How were the NTA/YTA labels attained?
Hi @nicoco404 ,
I'm interested in using this AITA dataset for some of my experiments, which I'm happy to credit to your dataset.
Would you be able to share more details on the dataset you've prepared? For example:
- What were the dates were these posts collected from?
- Did you perform any filtering on the posts?
- How were the NTA/YTA labels attained?
Thanks in advance!
--Jusitn
Hi Justin,
Feel free to use the dataset. You’ll find the code used to get the data from Reddit here as well: https://github.com/Nico404/AITA_data_exploration_and_ML
Data is all in a 2 week window around the time I pushed the code. No filtering occurred (except having a minimum number of comments for the ratings. I think 20 comments if I remember correctly) and I just counted the number of ratings in the comments for each one of them. There are also a few Jupyter notebooks.
Things I noted: Reddit only keeps around 2-week old fresh data that you can pull from the api. The rest is archived and not available. You can run the code multiple times with time delta if you need a bigger dataset. Also be mindful the proportion of YTA / NTA is not equal. Mapping of labels is also in the notebooks.
Cheers and happy to see what you’ll use it for.
Nicoco404
Hi @nicoco404 ,
I appreciate the reply. Thank you for the pointers!