Tools for Social Media Data [#]
Social media provides data that is highly valuable to many organizations, for example as a way to track public opinion about a company’s products or to discover attitudes towards “hot topics” and breaking news stories. However, processing social media text presents a set of unique challenges, and text processing tools designed to work on longer and more well-formed texts such as news articles tend to perform badly on social media. To obtain reasonable results on short, inconsistent and ungrammatical texts such as these requires tools that are speciﬁcally tuned to deal with them.
This chapter discusses the tools provided by GATE for use with social media data.
17.1 Tools for Twitter [#]
The Twitter plugin contains several tools useful for processing tweets. This plugin depends on the Stanford_CoreNLP plugin, which must be loaded ﬁrst. This includes tools to load documents into GATE from the JSON format provided by the Twitter APIs, a tokeniser and POS tagger tuned speciﬁcally for Tweets, a tool to split up multi-word hashtags, and an example named entity recognition application called TwitIE which demonstrates all these components working together.
17.1.1 Twitter JSON format [#]
Twitter provides APIs to search for Tweets according to various criteria, and to collect streams of Tweets in real-time. These APIs return the Tweets in a structured JSON format1 which includes the text of the Tweet plus a large amount of supporting metadata. The GATE Twitter plugin contains a format analyser for this JSON format which allows you to load a ﬁle of one or more JSON Tweets into a GATE document. Loading the plugin registers the document format with GATE, so that it will be automatically associated with ﬁles whose names end in “.json”; otherwise you need to specify text/x-json-twitter for the document mimeType parameter. This will work both when directly creating a single new GATE document and when populating a corpus.
Each tweet object’s text value is converted into the document content, which is covered with a Tweet annotations whose features represent (recursively when appropriate, using Map and List) all the other key-value pairs in the tweet object. Note: these recursive values are diﬃcult to work with in JAPE; the special corpus population tool described next allows important key-sequences to be “brought up” to the document content and the top level of the annotation features.
Multiple tweet objects in the same JSON ﬁle are separated by blank lines (which are not covered by Tweet annotations).
As well as the document format parser to load Tweets into a single GATE document, the plugin provides a “Populate from Twitter JSON ﬁles” option on the GATE Corpus right-click menu. Selecting this option bringgs up a dialog that allows you to select one or more ﬁles of tweets in the Twitter API’s JSON format and set the following options to populate the corpus.
- The default here is UTF-8 (regardless of your Java default) to conform to Twitter JSON.
- One document per tweet
- If this box is ticked (the default), each tweet will produce a separate document. If not, each input ﬁle will produce one GATE document.
- Content keys
- The values of these JSON keys are converted into strings and concatenated into each tweet’s document content. Colon-delimited strings specify nested keys, e.g., “user:name” will yield the value of the name key in the map that is the value of the user key. Missing key sequences are ignored. Each span of text will be covered by an annotation whose type is the key sequence.
- Feature keys
- The key sequences and values of these JSON keys (where present) are turned into feature names and values on the tweet’s main Tweet annotation.
- Save conﬁguration
- This button saves the current options in an XML ﬁle for re-use later.
- Load conﬁguration
- This button sets the options according to a saved XML conﬁguration.
Every tweet is covered by a Tweet annotation with features speciﬁed by the “feature keys” option. Multiple tweets in the same GATE document are separated by a blank line (two newlines).
Corpus population from Twitter JSON ﬁles is also accessible programmatically when this plugin is loaded, using the public static void method gate.corpora.twitter.Population.populateCorpus(final Corpus corpus, URL inputUrl, String encoding, List<String> contentKeys, List<String> featureKeys, int tweetsPerDoc).
17.2 Low-level PRs for Tweets [#]
The Twitter plugin provides a number of low-level language processing components that are speciﬁcally tuned to Twitter data.
The “Twitter Tokenizer” PR is a specialization of the ANNIE English Tokeniser for use with Tweets. There are a number of diﬀerences in the way this tokeniser divides up the text compared to the default ANNIE PR:
- URLs and abbreviations (such as “gr8” or “2day”) are treated as a single token.
- User mentions (@username) are two tokens, one for the @ and one for the username.
- Hashtags are likewise two tokens (the hash and the tag), but see below for another component that can split up multi-word hashtags.
- “Emoticons” such as :-D can be treated as a single token. This requires a gazetteer of emoticons to be run before the tokeniser, an example gazetteer is provided in the Twitter plugin. This gazetteer also normalises the emoticons to help with classiﬁcation, machine learning etc. For example, :-D, and 8D are both normalized to :D.
The “Tweet Normaliser” PR uses a spelling correction dictionary to correct mis-spellings and a Twitter-speciﬁc dictionary to expand common abbreviations and substitutions. It replaces the string feature on matching tokens with the normalised form, preserving the original string value in the origString feature.
The “Twitter POS Tagger” PR uses the Stanford Tagger (section 23.26) with a model trained on Tweets. The POS tagger can take advantage of expanded strings produced by the normaliser PR.
17.3 Handling multi-word hashtags [#]
When rendering a Tweet on the web, Twitter automatically converts contiguous sequences of alpha-numeric characters following a hash (#) into links to search for other Tweets that include the same string. Thus “hashtags” have rapidly become the de-facto standard way to mark a Tweet as relating to a particular theme, event, brand name, etc. Since hashtags cannot contain white space, it is common for users to form hashtags by running together a number of separate words, sometimes in “camel case” form but sometimes simply all in lower (or upper) case, for example “#worldgonemad” (as search queries on Twitter are not case-sensitive).
The “Hashtag Tokenizer” PR attempts to recover the original discrete words from such multi-word hashtags. It uses a large gazetteer of common English words, organization names, locations, etc. as well as slang words and contractions without the use of apostrophes (since hashtags are alphanumeric, words like “wouldn’t” tend to be expressed as “wouldnt” without the apostrophe). Camel-cased hashtags (#CamelCasedHashtag) are split at case changes.
More details, and an example usecase, can be found in [Maynard & Greenwood 14].
17.4 The TwitIE Pipeline [#]
The Twitter plugin includes a sample ready-made application called TwitIE, which combines the PRs described above with additional resources borrowed from ANNIE and the TextCat language identiﬁcation PR to produce a general-purpose named entity recognition pipeline for use with Tweets. TwitIE includes the following components:
- Annotation Set Transfer to transfer Tweet annotations from the Original markups annotation set. For documents loaded using the JSON document format or corpus population logic, this means that each Tweet will be covered by a separate Tweet annotation in the ﬁnal output of TwitIE.
- Language identiﬁcation PR (see section 15.1) using language models trained on English, French, German, Dutch and Spanish Tweets. This creates a feature lang on each Tweet annotation giving the detected language.
- Twitter tokenizer described above, including a gazetteer of emoticons.
- Hashtag tokenizer to split up hashtags consisting of multiple words.
- The standard ANNIE gazetteer and sentence splitter.
- Normaliser and POS tagger described above.
- Named entity JAPE grammars, based largely on the ANNIE defaults but with some customizations.
Full details of the TwitIE pipeline can be found in [Bontcheva et al. 13].