Twitter Sentiment Tracking for Predicting Marketing Trends - Smart Information Systems: Computational Intelligence for Real-Life Applications - page 57

language data. The system offers interfaces that are not difficult to use to more than
50 corpora and lexical resources including WordNet. Moreover, the system also
provides a set of text processing libraries for classification, tokenization, stemming,
tagging, parsing, and semantic reasoning.
Regarding further benefits of the system, Loper and Bird [ 19 ] stated that NLTK
leads to a simple, extensible, and even framework for projects and assignments. They
declared that NLTK is well documented, easy to understand how it works, and simple
to use.
In this chapter, NLTK is used to classify the tweets by using the Naïve Bayes and
Maximum Entropy classifier methods.
2.3.1.3 Twitter API
Twitter API is described, by Sharifi et al. [ 29 ], as an API based completely on HTTP,
and it is provided by Twitter. With Twitter API, users can accomplish nearly any
task that can be achieved through Twitter's web interface. As for the nonwhite listed
users, Twitter Rest API allocates 150 requests per hour to a user.
Fortunately, Twitter Search API, which is used in this chapter, does not have this
kind of a restriction for developers. But frequency and complexity of requests is
important to avoid being in blacklisted users.
Certain points are found crucial to be grasped before using the Search API. For
example, the Search API is an index composed of the most recent tweets, not an
index demonstrating all tweets. Currently, the index incorporates tweets of 6-9 days.
Furthermore, the Search API cannot be used to search for tweets that are older than a
week. Queries are subject to restrictions owing to complexity. In this case, the Search
API will report an error as a response. All queries are made without identification
to be provided; in other words, search does not require authentication. The search
pays attention to relevance, not to completeness. This may result in some tweets and
users' being missed from the search results. The Search API cannot use the near
operator, so the geo-code parameter should be used. Queries are restricted to 1,000
characters, including any operators. During the process of geo-based searches with a
radius, 1,000 different sub-regions will be taken into consideration when evaluating
and processing the query.
In this chapter, Twitter Search API is used to collect tweets.
2.3.1.4 Google Maps API
The Google Maps API, which is a free service provided by Google, allows devel-
opers to embed high-resolution maps into their web pages by using the JavaScript
technology.
Furthermore, the API provides various functions that enable manipulation of the
maps as well as making it possible to make additions to the content of the map via
lots of services. Using this API, the users are enabled to design and create strong
maps applications on their websites.