In the past few years, research on social media has emerged as an important area as they are playing a leading role in society. Social media platforms help to forecast or predict events, using thoughts of people by collecting, storing and analyzing data for the purpose of harvesting information
Social media analytics is focused with ,developing and evaluating informatics tools and frameworks to collect, monitor, analyze, summarize, and visualize social media data to facilitate conversations and interactions to extract useful patterns and intelligence” . Therefore, the development of effective and efficient analytics techniques for social media analysis becomes essential. To conduct social media analytics, data mining, text analysis and related advanced analytics techniques, e.g., sentiment analysis and semantic analysis techniques are frequently adopted. Recently there has been strong interest in the power of social media analytics in creating new value, supporting decision making and enhancing competitive advantage. A number of studies from various research communities have been devoted to unveil the value, impact and implications of social media analytics.
During my internship at WS02 I worked on a project on social media analytics to solve real time analysis on tweets. The social media that we chose is Twitter as it is micro blogging service that has been become popular due to its data volume, availability in mobile devices and the real time nature and avalanche effect caused due to re-tweeting makes it an informative source for potentially changing trends. This twitter analytics is a solution designed by me with the WSO2 Stream Processor. This solution is by default, shipped with the Stream processor for every release.Twitter Analytics solution allows you to create a Twitter application that can be integrated to your service, and monitor the Tweets generated from it.The output of this project is a dashboard of different widgets in multiple pages. This dashboard contains ten widgets.
1. Hashtag Widget: This widget will display the name of the hash tag that we expect to do analysis. This hash tag is read from the siddhi application which is deployed in worker/editor run time.
2. TweetsCounter Widget: This widget will show the real time count of tweets since the dashboard is started. If the dashboard is refreshed, then it will again count from zero. For this I used Publisher-Subscriber concept to count the live tweets. This widget is a subscriber and its Publisher is LiveTweets widgets.
3. TweetCountAnalysis Widget: This widget contains 2 types of graphs, such as last hour graph and last 24hrs graph. For this widget I have used Siddhi incremental aggregation for grouping data by seconds, minutes, hours, months and years.
4. LiveTweets Widget: When we start the dashboard, this widget will show the last 5 tweets which are stored in the database. Once the Siddhi application is started, it will show most recent tweet at the top of the widget. To display tweet, I use some npm libraries. Such as : react-tweet-embed , react-custom-scrollbars
5. Emotions Widget: This widget will illustrate how people react (positive, negative, neutral) about that particular hash tag over 24 hours. To find out each tweet’s emotion type by using ‘sentiment extension’ which reads each word from the tweet and calculate its sentimental value. Then I determined emotion type of each tweet.
6. EmotionsAnalysis Widget: This widget expresses a graph of sentimental value for the last hour. We need this widget to express how people react per minute.
7. TopSentiment Widget: This widget will show the tweets, which is in descending order of its sentiment value. The tweet which has top most sentiment value will be in the top of the widget.
8. WordCloud Widget: This widget contains 3 types of word clouds. They are text cloud, hash tag cloud and mention cloud. The size of the word will change according to the frequency of that word. For this widget I had used a npm library. That is react-tag-cloud.
· Text Cloud: This is a cloud of texts that are used in tweets.
· Hashtag Cloud: This is a cloud of hash tags that are used in tweets.
· Mention Cloud: This is a cloud of mentioned names that are used in tweets.
9. TopCountries Widget: This widget will show the tweet count of each country for last 24 hours when we click on or hover over a country.
10. PopularTweets: This widget displays the most popular tweets with that particular hashtag.
To monitor Twitter activity via the Twitter Analytics dashboard, follow the prerequisites and procedure that you need to do , for that refers