Trend square: An Android Application for Extracting Twitter Trends Based on Location

摘要

The identification of popular and important topics discussed in social networks is crucial for a better understanding of societal concern. Without having to explore the vast amount of information, it helps users to be aware of the trends around them. Trend detection methods introduced so far have not used the location matching feature thus lack integrating geographic locations information with the analyzed trend. To address this problem, this research aims at developing an Android mobile application, TrendSquare which uses LDA topic modeling algorithm to obtain location-specific trending tweets. This is not an exhaustive trend analysis of all services, because of the vast amount of tweets present; rather it is for picked services (Restaurants, Jewelry, Shopping malls, Textiles) where trend notifications to customers would serve as a prominent way of advertising and target marketing. An outstanding advantage of using LDA for Twitter trend detection is that it can well capture events with narrow topical scope. This implementation also provides us favorable results in the accuracy of trends obtained.