Epinomy

One of the biggest challenges of enterprise search is the variety and number of sources where data resides. Managing all those sources can be a nightmare.  Epinomy provides a point & click dashboard that makes ingesting data from almost any source a breeze.  Structured, unstructured, web feeds, document folders, data in other repositories…ingest it or connect to a source where the data resides elsewhere using the Data Source Registry built into Epinomy.

Features

• List of Feeds
  
  The Data Source Registry displays a comprehensive list of all the sources of data searchable in Epinomy. Click on one of the sources to manage it in Epinomy.

• Connect
  
  Connecting a source to Epinomy simple. Define the source and where it lives and Epinomy does the work.  Epinomy can connect to multiple data sources and ingest multiple data formats into it's three fundamental internal data structures: tables, text and triples.

  • Tables - Tables are structured data sets, particularly time series.  Tables contain the structured data in an enterprise, including:

    • Time Series

    • Financial Data

    • System Logs

    • Telematics

    • Signals

    • Internet of Things

    • Purchase Behavior

    • Customer Metrics

    • Query History

  • Text - Text is unstructured data in the form of documents.  Some examples of text includes:

    • E-Mail

    • Office Documents

      • Word

      • Excel

      • Powerpoint

    • PDF

    • XML Documents

  • Triples - Semantic triples are the best way to represent linked structured data including:

    • RDF

    • Vocabularies

    • SKOS (Simple Knowledge Organization System)

    • Taxonomies

    • Ontologies

    • Linked Open Data

    • Concepts

    • Social Graphs

    • Site Maps

• Load
  
  Once Epinomy has connected to the source, describe what you want loaded and how you want it loaded.

• Transform
  
  Epinomy transforms the data from individual sources by running any number of processes that will format the data being ingested from that source that will define how that data interacts with other feeds and performs in search.
  
  Some examples of pre-configured pipeline stages include:

  • Dimensional Analyser - extracts concepts from structured data to be stored in the triple store
  • Element Mapper - Maps input elements and columns to normalized internal format.
  • Disambiguator - Automatically creates links among similar concepts. 

• Schedule
  
  The scheduler permits periodic polling of data sources for new information.  Depending on the currency of the data source, periodicities can vary from real-time to one-time.