Mr
Sebastian Schlag
(ITI, KIT)
10/11/17, 2:00 PM
Presentation
In computer science, engineering, and related fields graph partitioning (GP) is a common
technique for processing very large graphs, e.g. networks
stemming from finite element methods, route planning, social networks or web graphs.
For example, in parallel computing good partitionings of unstructured graphs
are very valuable. In this area, graph partitioning is mostly used to partition the...
Erhard Rahm
(Universität Leipzig),
Eric Peukert
(Universität Leipzig)
10/11/17, 2:30 PM
Presentation
Die Analyse sehr großer Netzwerkdaten gewinnt immer mehr an Bedeutung, zum Beispiel um Erkenntnisse aus Logistik-, Geschäfts- und sozialen Netzwerken zu gewinnen. Durch die Repräsentation von Netzwerkdaten als Graph lassen sich komplexe Beziehungsgeflechte zwischen heterogenen Datenobjekten analysieren. In der Forschung existieren bereits wertvolle analytische Graph-Algorithmen, die jedoch oft...
Dr
Martin Skorsky
(Software AG)
10/11/17, 3:00 PM
Demonstration
This demonstration will show the platform for streaming analytics of social media data we developed for iTESA. iTESA is one of the Smart Data projects funded by the BMWi. Main components of this platform are Software AG's Apama for streaming analytics and Kafka and Flink at Fraunhofer IVI for Dynamic Semantic Data Mining and Fuzzy Association Rule Mining.
Focus of this demonstration is...
Mr
Jonas Traub
(Technische Universität Berlin)
10/11/17, 4:00 PM
Presentation
We present two research works dealing with massive sensor data inputs.
1) We present I², an interactive development environment for real-time analysis pipelines, which is based on Apache Flink and Apache Zeppelin. The sheer amount of available streaming data frequently makes it impossible to visualize all data points at the same time. I² coordinates running cluster applications and...
Mr
Johannes Luong
(Technische Universität Dresden)
10/11/17, 4:30 PM
Demonstration
# AL in action: unified relational- and graph processing
Modern data analyses frequently involve a variety of data types and corresponding programming models.
Popular big data platforms, such as Apache Hadoop, Spark, or Flink, hide this variety behind a generic and sometimes system oriented programming abstraction.
In recent years, domain specific languages that sit on top of those low...
Mr
Jan Frenzel
(TU Dresden)
10/11/17, 5:00 PM
Demonstration
In the era of ubiquitous data, especially coming from sensors of all kind providing time-dependent data, applying analytics methods to characterize the value in the observed processes is a promising, but also challenging task. Not just the sheer amount of data, but also the integration and verification of data at hand needs to be handled prior to any analysis of time series data. As a general...
