Analytics Brainstorming

Stakeholders

• Internal:
  • how do people use chaise
  • improve usability
  • is faceted search working
• Funders:
  • how much activity for a funded data hub
  • how many visits to a site
  • bounce rate
  • where are people coming from (geo)
  • referrals from other sites
  • overall data downloads
  • registered users, (average) time spent on site, frequency of visits
• Hubs/Consortiums:
  • provide feedback for data producers
  • page visits for specific data
  • downloads of specific data
  • uploads of data

Usage Events

Searching:

• who is logging in
• how long they are staying
• browse without logging in
• downloads
• how do users use filters,
  • e.g., do they search first, then add filters,
  • do they explore the available attributes ("all attributes")
  • specifically, which attributes do they tend to use (per project/catalog basis)
  • what values do they set in filters
• what search terms are used
• do users start their chaise/search session with or without filters set (i.e., do they start from scratch or do they often use a bookmark that has filters set in it, for example from a project home page or from a bookmark shared by another user or saved by themselves)
• delete an individual in-use filter, or do they 'clear all'
• use of layouts (card vs table vs list)
• use of sorting
• do they copy a permalink

Details:

• do they use Previews
• do they use other iFrame content
• do they expand nested tables
• do they leave using "external references" links, or pubmed, geo, etc. links
• do they click on links to related data (e.g., links that take them back to the search page)
• how can we tell if they find what they were searching/browsing for? maybe terminate in a download or browsing a 3D image etc
• do they click on links to vocabulary terms (e.g., mouse age stage)

Tactically

• continue using Google Analytics (open to debate)
• and Google Tag Manager
• if using GA, need to be able to configure separate accounts for each deployment (don't use one master GA account across all projects)
• check on the free tier and cost for GA (answer: 10M hits per month)
• Angulartics (see http://angulartics.github.io)
  • this is a wrapper API around various analytics providers and integrates with Angular framework
• add custom event types, and integrate event tracking in angular apps
• do we need to define a set of events? tags vs events?