SkyKeeper is a powerful and user-friendly platform that has the ability to serve sentiment analysis services, crisis analysis, and success analysis by completely running on cloud systems to crawl big data and to provide results to thousands of users in small amount of time.
Since the development of big data and cloud computing technologies, SkyKeeper offers a more meaningful and detailed social media analysis to companies and institutions by using scientific methods. SkyKeeper has a user interface that exposes analysis’ results by using interactive visuals.
How does Skykeeper analyze the data?
Keywords are required for analyzing shared contents on social media (mined from Twitter, Facebook, etc.) and data is collected by using these keywords.
After filtering personal data from collected social media data, the filtered data is archived. Then that filtered data is processed by categorization and analysis through stream data processor frameworks (such as Storm) and saved to NoSQL database in a queryable format which improves the performance of the analysis for future purposes.
The fast and continuous, sentiment analysis process is started by using distributed, natural language processing techniques with Hadoop on saved data. Intermediate sentiment analysis results are generated by considering time, geographical region and user demands via Hive which is a SQL to Map/Reduce conversion based framework. The final analysis results are generated by processing these intermediate results with right evaluation metrics.
These analyzed results are saved in accessable and queryable format on cloud. Associated users are informed with pre-determined ways (such as E-Mail, SMS, etc.). Users can access the analyzed results via mobile and web interface easily. SkyKeeper offers fascinating and drill down tools on reports generated by users’ demands.
• Analyzes social media in real time. Provides detailed and location-based social media analyses.
• Archives all Turkish social media, so that historical analysis can be performed.
• Performs natural language processing and analyzing in Turkish, English, Spanish and Arabic Languages.
• Currently, Twitter and Foursquare data are processed. If required, other social media platforms can be processed.
• Archived data is processed using Map/Reduce (Hadoop) and natural language processing. Big data is processed in a distributed manner.
• Sentiment of Turkish, English, Arabic and Spanish data in social media is analyzed via Sentiment Analysis methods.
• Offers analyses of social trends, user behavior, product satisfaction and brand preferences.
• Clusters of human are identified in real time via location information.
• Runs on cloud. You pay as much as you use and your services can be scaled. Also Skykeeper can be deployed on a private cloud for special customers.
• Provides easy and effective access to the reports of the analyses via user-friendly mobile and Web interfaces.
• Contains alarm mechanisms that warn users and other systems via mail or other media in predefined situations.
• Can be used in several different domains, such as “Cyber Security”.