Logo des Repositoriums
 

Rightinsight: open source architecture for data science

dc.contributor.authorBulut, Ahmet
dc.contributor.editorRitter, Norbert
dc.contributor.editorHenrich, Andreas
dc.contributor.editorLehner, Wolfgang
dc.contributor.editorThor, Andreas
dc.contributor.editorFriedrich, Steffen
dc.contributor.editorWingerath, Wolfram
dc.date.accessioned2017-06-30T11:39:35Z
dc.date.available2017-06-30T11:39:35Z
dc.date.issued2015
dc.description.abstractWe give the details of our reference architecture called RightInsight for enabling rapid data science. RightInsight is based purely on open source technologies. The data is stored in a standard distributed file system such as HDFS. The stored data is processed in Apache Spark, which provides an enhanced Map/Reduce programming environment. Its rich and powerful machine learning base makes it easy to construct descriptive, prescriptive, and predictive models. In addition to providing an agile environment for making sense of the data and the data science problem at hand, its Python-based middleware with a wide array of scientific libraries such as scipy, numpy, matplotlib, and pandas, enables interactive and exploratory data analysis. The ability to ask questions, especially the right questions, and to do what-if analysis is extremely important for any serious data science project. The results of such exploratory analyses are stored in a suitable format that is easily consumable in the Web tier. Using rich JavaScript libraries such as data driven documents and bootstrap, the formatted data can be visualised within a Web browser in creative ways for rapid insight discovery.en
dc.identifier.isbn978-3-88579-636-7
dc.identifier.pissn1617-5468
dc.language.isoen
dc.publisherGesellschaft für Informatik e.V.
dc.relation.ispartofDatenbanksysteme für Business, Technologie und Web (BTW 2015) - Workshopband
dc.relation.ispartofseriesLecture Notes in Informatics (LNI) - Proceedings, Volume P-242
dc.titleRightinsight: open source architecture for data scienceen
dc.typeText/Conference Paper
gi.citation.endPage160
gi.citation.publisherPlaceBonn
gi.citation.startPage151
gi.conference.date2.-3. März 2015
gi.conference.locationHamburg

Dateien

Originalbündel
1 - 1 von 1
Lade...
Vorschaubild
Name:
151.pdf
Größe:
543.11 KB
Format:
Adobe Portable Document Format