Deep Email Miner

Features

The most important feature of the Deep Email Miner v1 is the combination of network, email and text analysis under a common user interface.
Social Network Analysis
  • Create social network with employees as nodes
  • Edges indicate email traffic between two employees
  • Find clusters of employees
  • Rank nodes/edges on centrality measures
  • Graph statistics
Email Analysis
  • Find email threads
  • Display emails
  • Add and remove network filters based on different email attributes and the number of emails sent
  • Email corpus statistics
Text Mining
  • Label emails as business/private
  • Create word list
  • Create word vectors
Input/Output
  • Load network and single emails from database
  • Cache email bodies to balance the amount of database-interaction and the program's memory usage
  • Load/Save email labels to file
  • Export word vectors to .csv
  • Export network to Pajek .net
Software Features
  • Efficient enough to handle email corpora with several hundred people
  • Written in Java, runnable everywhere
  • Additional version as Win32 executable
  • Configurable with properties-file