Admin Setup

Kibana—Analytics Data Visualization Workbench

Available in:

  • Ultimate and Enterprise editions
    Get in touch with Sales for licensing questions.

Memsource users are able to visualize data from their accounts through Kibana, an open-source data visualization workbench. Kibana allows you to visualize and explore all data indexed by our analytics data warehouse, and there is an almost endless number of queries that can be visualized using Kibana.

Data Types That Can be Visualized

All visualizations created in Kibana can also be turned into CSV based tables containing the hard data behind. Here are some examples of data indexed through Kibana:

  • Project
    • Project creation date, last modification, created by, last modified by, source language, target languages, project settings, client, domain, subdomain, business unit, project status, due date, etc.
  • Jobs
    • Job creation date, last modification, created by, last modified by, source language, target language, assigned to linguist/vendor, due date, workflow step, file type, etc.
  • Analyses
    • Total wordcount, net wordcount, translation memory matches, machine translation matches, etc.
  • Costs
    • Costs per project, project margin, etc. (based on Price Lists).
  • API
    • Summary of API usage, list and breakdown of API calls per period of time, etc.
  • Time Tracking
    Collection of time-tracking stats (in ms) follows these principles and can be seen in both MXLIFF files and in Kibana:
    • User clicks into segment #1:
      • Counting of segment #1 thinking time starts
      • Counting of segment #1 editing time starts
    • User starts editing segment #1:
      • Counting of segment #1 thinking time is stopped
      • Counting of segment #1 editing time continues
    • User clicks into segment #2:
      • Counting of segment #1 editing time is stopped
    • When a user re-enters segment #1, the counting starts again and the new times are added to the previous values. If edited upon re-entering, both times are updated. If there is no editing, times are not updated.
    • The thinking and/or editing times are measured for each workflow step separately.
Note: Kibana version 4.1.11 is required to process data from Memsource. Download from the Elastic website

Installation

Memsource steps

  1. From the Setup page, click Kibana under Analytics and Reporting.
    The Kibana setup page opens.
  2. Click Setup Kibana Access.
    A unique URL is generated that required to visualize data with Kibana.

Kibana steps

  1. Open the kibana.yml configuration file from Kibana config subfolder. 
  2. Replace elasticsearch_url value (usually "http://localhost:9200") with the URL generated in the Memsource steps.
  3. Ensure the kibana_index value is set to .kibana
  4. Save the kibana.yml file.
  5. Start Kibana from the the Kibana bin subdirectory.
    • Windows: double-click kibana.bat
    • Other O/S: double-click kibana (other users).
  6. Verify setup.
    The following is visible in the command prompt window:

    Important: Command prompt window must stay open to maintain connection to Kibana.
  7. Enter localhost:5601 into the address bar of a browser to access the visualization workbench.

Kibana Configuration

  • Index patterns are required for visualization and are applied in Settings under the Indices menu. The following index patterns are currently available:
    • com.memsource.web.analytics.data.job (job.dateCreated)
    • com.memsource.web.analytics.data.jobpart (jobPart.dateCreated)
    • com.memsource.web.analytics.data.analysis (jobPart.dateCreated)
    • com.memsource.web.analytics.data.costs (jobPart.dateCreated)
    • com.memsource.web.analytics.data.api (request.date)
    • com.memsource.web.analytics.data.lqa (lqa.dateCreated)
  • When creating indexes, the expression in brackets represent the recommended Time-field name for the given index. Any or all of the indexes available can be configured.
  • Once the indexes are registered, data can be visualized through the Visualize tab.

Index Classifications

  • Job index
    • Contains information about job sizes (words, pages etc.), who created the jobs, when they were created, filetypes, projects, and organizations they belong to.
  • Job Part index
    • Jobs themselves do not have statuses or assigned users. This type of information is stored in terms of job parts (i.e. jobs in different workflow steps constitute different job parts). Otherwise, the same of data as in the Job index are available (filetype, project, organization etc.).
  • Analysis index
    • Information about Analyzed Job Parts (divided into match categories) with and without Net Rate applied.
  • Costs index
    • Costs for the Job Part (based on Price Lists).
  • API index
    • Summary of API usage.
  • LQA index
    • Breakdown of LQA evaluations entered for Job Parts based on defined categories, subcategories and severity.
Tip: A selection of configuration samples
Was this article helpful?

Comments

0 comments

Please sign in to leave a comment.