Social media analysis with FKIE and CONET

How can the Internet and how can social media in particular be monitored? How is it possible to find out relevant information, incorrect information, opinions and trends in a timely manner and to analyze them in relation to one’s own needs? With NewsHawk®, Fraunhofer FKIE has developed a tool for this task. With a platform based on this, CONET plans to make the results of the researchers usable for customers in all areas. In this blog post, we outline what can result when state-of-the-art research is combined with a service-oriented way of thinking.

The platform in detail: From data acquisition to data visualization

The targeted platform can collect, record, analyze and visualize content relevant to the user. It will be great for monitoring social media. The platform can be used in a variety of ways for various questions. It consists of three building blocks:

• data collection: In the first step, relevant posts are recognized and stored in a smart database. Each user can set their interests individually using keywords that can be prioritized, regularly linked and updated at any time.
• data evaluation: In the second step, the platform processes the contributions statistically and analytically. Connections are worked out using various methods, such as statistical processes, natural language processing (NLP), but also machine learning / deep learning.
• data visualization: In the third step, the recognized results, connections and contexts that contribute to a complex of topics – such as communication networks between social media users – are visualized in a user-friendly way. Using the interactive, graphical user interface, even inexperienced users can quickly identify information and read the results relevant to their questions. This is guaranteed by the consistent implementation of ergonomic principles.

Depending on the needs and the type of analysis question, different analysis modules are combined and interconnected. The modules are suitable for both online and complex offline analyses.

Application areas of the analysis platform

The analysis platform is a flexible tool and basically suitable for all applications where:

• unstructured or semi-structured large amounts of text present,
• there are relationships between the individual texts that are not directly recognizable,
• in large amounts of data connections and abnormalities should be specifically identified.

through the Cooperation of CONET with Fraunhofer FKIE The latest scientific findings and processes, for example in the areas of NLP and machine learning, which are NewsHawk® proven, adapted and are thus available to users of the platform.

Example: social media analysis

A central application of the analysis platform is the Observation of social networks (specifically Twitter). After entering possible keywords for the first time, the relevant keywords and hashtags are quickly recognized so that the keywords are adapted. After that, real-time monitoring of the social networks can be used to identify the following aspects, for example:

• Topics are discussed more intensively
• certain networks of users occupy a subject area
• Influencing the topic by individual users (influencers) within the social network
• Shifting the topic of discussions (e.g. from positive to negative perception)
• Identification of important (influential) individual tweets

The platform will run on a central server and be addressed via a web interface. With the appropriate access rights, the tool can be used and configured individually. The example shows how CONET combines process and technology know-how with many years of project experience. This opens up unexpected – and sometimes unusual – potential and opportunities for our customers.

Was this article helpful to you? Or do you have any questions? Write us a comment or give us a call.

Link Tips