Q&A Software Embedded Data Analytics

4 minute read

To gather information about the usage details of our Q&A software, we use a two-tier approach to software embedded data analytics. It is based on an ELK (Elasticsearch/Logstash/Kibana) stack on one side and Python-based Jupyter notebooks on the other side.

The business knowledge captured through the use of an in-house Q&A system extends well beyond the facts and figures contained in the submitted questions and answers.

Beyond the content (i.e. questions, answers and comments) stored within the system, lots of metadata is simultaneously generated.

This metadata can provide business analysts, knowledge managers and HR staff with valuable insights regarding

  • knowledge hubs, i.e. individual users or user groups with particular expertise
  • knowledge gaps, i.e. issues for which internal expertise is limited
  • employee engagement, i.e. speed and extent of system adoption, support among employees,…
  • company culture, i.e. openness to sharing information, collaboration, etiquette,…
  • “superstars”, i.e. highly engaged and knowledgeable employees

From the perspective of the software developer, software embedded data analytics furthermore provides useful information for improving and extending the products and services offered to customers.

These include i.a. the time users spend navigating or searching, the time users need for completing particular tasks within the system and the ease of integration with additional software.

At 42ways, we opted for a two-tier approach to software embedded data analytics.

At a basic level, ELK Stack-based collection, visualization and evaluation of log files allows us to extract metadata from our user’s interaction with the Q&A software. Thereby, we generate insights primarily relevant to the continuous improvement of our product, i.e. information relating to use times, navigation and search.

Recently, we additionally started implementing high-level Python-based exploratory data analysis facilitated through Jupyter notebooks. Through this advanced level we explore, which data would be useful for our customers in order for 42ways Q&A to become a tool for business intelligence.

To this end, we are currently planning to steadily integrate Python-based dashboards into a user-friendly interface to enable non-tech users (e.g. community managers, knowledge managers, …) access to the wealth of data underlying our Q&A software.