obiba/mica

Publish with Mica

Web data portals to describe studies, networks and variables, with real-time Opal queries.

DocumentationGitHubDemo

What is Mica?

Mica creates data web portals for large-scale epidemiological studies and multiple-study consortia.

Mica gives studies scientifically robust data visibility and a web presence without significant IT effort: structured descriptions of networks, studies and harmonization initiatives, annotated and searchable data dictionaries, and data access request management. Mica stores no record-level data: it works with Opal for the data dictionaries and summary statistics, and with Agate for the users.

Features

Catalogue. Documents specific to the epidemiological domain, with their relationships:

  • Networks of studies.
  • Individual studies, with populations and data collection events, and harmonization initiatives.
  • Collected datasets and harmonization protocols, with their variables, data schema and harmonized variables.
  • Research projects conducted with the data.

The description forms are customizable, and documents are annotated with taxonomies (such as the Maelstrom Research classification), either manually or inferred from their variables.

Search. A faceted search engine, that combines criteria across document types: for instance, variables about alcohol in studies that collected biosamples and belong to a network. Variables, studies and networks can be collected in a cart, compared, and downloaded as CSV.

Data summaries. For studies using Opal as their secured database, a distributed query engine gives authorized researchers real-time aggregated reports on the number of participants with specific characteristics, and summary statistics on each variable. No individual-level data passes through Mica. Data dictionaries can also be provided as Excel or SPSS files, without an Opal server.

Data access management. Electronic handling of access requests, from submission to approval:

  • Forms designed by the administrator: main application, preliminary, feasibility, amendment and end-user agreement forms, with attached files and a selection of requested variables.
  • Configurable workflow (review and conditional approval steps), with email notifications at each status change.
  • Collaborators, private comments for the data access committee, and a history of each request.
  • Research projects reporting what was done with the data.

Publication flow. Every document and file goes through draft, review and publication states, with permissions separating editors from reviewers, revision history and comments.

Web portal. The public pages are FreeMarker templates that can be customized, extended or overridden, with a theme based on AdminLTE, custom pages, translations and multilingual content.

Security. Authentication delegated to Agate (two-factor authentication, OpenID Connect realms, roles mapped from Agate groups), permissions on any document, sign-up protected by Google reCAPTCHA.

Automation. The micar R package, the obiba-mica Python client and the REST web services allow exploring the published content and scripting administration tasks. Search engines (Elasticsearch, OpenSearch), table and taxonomy sources are plugins.

Download

Mica is a stand-alone Java web server application. The Debian and RPM packages are strongly recommended, as they greatly simplify installation and upgrades.

See the installation guide and the configuration guide.

Try Mica

Have a look at the demo site.

Administration pages:

  • Administrator: administrator / password
  • Content editor: editor / P@ssw0rd
  • Content reviewer: reviewer / P@ssw0rd

Data access pages:

  • Data access officer: dao / P@ssw0rd
  • Regular user: user1 / P@ssw0rd

Mica releases