This repo hosts the post data retrieved from Dcard API, which were colleceted for the purpose of building a small corpus. These posts came from the top-100 popular forums of Dcard. Each post is at least 100-character-long.

The post data were segmented and PoS tagged using ckiplab/ckiptagger.

• data/dcard.jsonl: The segmented and tagged corpus. Each line is a json string representing a post.
• data/rawdata.zip: The raw data retrieved from https://www.dcard.tw/_api/forums and https://www.dcard.tw/_api/posts.

The quickest way to query KWIC concordance in this corpus with this concordancer is using docker.

Download image:

docker pull liao961120/dcard

Run server:

docker run -it -p 127.0.0.1:1420:80 liao961120/dcard

When you see Corpus Loaded printed on the command line, you can visit https://kwic.yongfu.name to use the app.

The source code of the concordancer can be found in liao961120/kwic and liao961120/kwic-backend. Read more about the concordancer in this post.

• Number of tokens: 5292615
• Number of posts: 19224
  • Female author: 12007 (62.46%)
  • Male author: 7217 (37.54%)