A key challenge of identifying sentiment in user-generated content is informal language use, since sentiment is often expressed by slang words. In order to tackle the challenge, we built and release a large scale sentiment lexicon of slang words, SlangSD, which contains over 90,000 slang words/phrases. SlangSD can be useful for research related to analyzing sentiment on social media platforms. The slang words are obtained from UrbanDictionary. The annotation of sentiment polarity is automatically conducted with the help of existing sentiment lexicons, content from Twitter, and the synonym feature of UrbanDictionary. A detailed introduction of how we built SlangSD, and preliminary results of leveraging it on sentiment analysis can be found here.

In the data files, each row corresponds to a slang word-sentiment polarity pair. The first column is a lowercased slang word, and the second column is the corresponding sentiment polarity. The two columns are split by a tab character ('\t').

Below are two rows from the lexicon:


perpindiculous -2

slap it fresh 1


The sentiment polarity takes 5 values: -2 (strongly negative), -1 (negative), 0 (neutral), 1 (positive), and 2 (strongly positive).


Please download SlangSD from here.


Please refer to this paper for the details of our functions:

Liang Wu, Fred Morstatter, and Huan Liu, SlangSD: Building and Using a Sentiment Dictionary of Slang Words for Short-Text Sentiment Classification

If you use it to analyze your data, please cite the above paper using this bib file:

  author    = {Liang Wu and Fred Morstatter and Huan Liu},
  title     = {SlangSD: Building and Using a Sentiment Dictionary of Slang Words for Short-Text Sentiment Classification},
  journal   = {CoRR},
  volume    = {abs/1168.1058},
  year      = {2016},
  url       = {http://arxiv.org/abs/1608.05129},
  timestamp = {Wed, 17 Aug 2016 23:32:57 GMT}

About us

Liang Wu (Student, ASU)
Fred Morstatter (Student, ASU)
Huan Liu (Professor, ASU)

If you have any questions or suggestions, please kindly let us know.

Existing Sentiment Lexicons
Web Statistics