metadata
license: apache-2.0
language:
- en
- zh
TLDR
- website: theedgemalaysia
- num. of webpages scraped: 414,268 (only webpages with full articles)
- link to dataset: https://huggingface.co/datasets/wanadzhar913/crawl-theedgemalaysia
- last date of scraping: 14th August 2023
- status: complete
- pull request: https://github.com/huseinzol05/malaysian-dataset/pull/261
- contributed to: https://github.com/huseinzol05/malaysian-dataset
Note
The "language" column for the data set has errors as it miscategorizes articles in the Mandarin language. This is primarily because I was searching for the string "English version" in the text. This will need to be accounted for if type of language used is important.
Methodology
For The Edge Malaysia, each of their articles seem to have a unique ID at the end of the url e.g., "677590" in "https://theedgemalaysia.com/node/677590". Hence, since we won't be able to do this by month, page no., etc., we'll use a brute force approach that tests every combination of numbers, such that we'll only scrape from a valid url.
Progress
- batch1
- batch2
- batch3
- batch4
- batch5
- batch6
- batch7
- batch8
- batch9
- batch10
- batch11
- batch12
- batch13
- batch14