FEATURES
Automatic connection and monitoring of public content
Automatic connection and monitoring of the public content of social sources in streaming like Twitter. Intermediate storage of the original tweets, always keeping links to the original sources for the generation of reports and their later verification.
Deep search engine
SAMi2 reuses the most common search ideas (search by users, by keywords) and by adding advanced searches (search by location, by related concepts, by times in which tweets and in which the aforementioned actions are performed).
Visual and dynamic interaction systems
Visual and dynamic interaction systems that facilitate the process of generated information: heat maps, hashtags clouds, social graphs
Artificial Intelligence
Refinement with artificial intelligence, which adapts future results to the needs of final users through simple and quick training
The Future
-
New Social Networks
Connecting to new social networks with new paradigms: Instagram (multimedia analysis), Facebook (social graph analysis) and other information sources such as Forums or Blogs (emphasis on static content).
-
Multimedia Analysis
Conversion of audiovisual content that users link to an additional source of information: people detection, scene description, optical character recognition in images.
-
Social Analysis
For interesting people detection, opinion leaders or content generators, SAMi2 provides visual tools for the registration of mentioned users, users relationship, traffic volume between those users and other social metrics.
USED TECHNOLOGIES
Big Data
High-performance of mass-production (Big Data) technologies with scalable Apache Spark deployments to provide the required data volume.
Natural Language Processing
Natural language processing with the Stanford NLP base stack, complemented with HI Iberia’s own tools for the Spanish language and, in particular, the truncated and simplified language that is used in Twitter: detection of synonyms, expansion of abbreviations and regularization of terms.
Artificial Vision
Artificial vision using convolutional neural networks (Deep Learning): scene description, people detection and text recognition in images (OCR).
Neural Networks
Refining with neural networks of results training of the system. Through a very agile process of classification of results by the user, SAMi2 learns from the user the tweets that are most relevant to prioritize them.