IEEE Computational Intelligence Magazine - November 2020 - 53

secure a data access agreement (DAA) in compliance with relevant ethics and privacy legislation.
The core data platform is based on Apache HDFS, Apache
Spark, and Apache Hive. In the system architecture diagram in
Figure 2, the light grey boxes indicate user-facing web applications, whilst the dark grey boxes represent services and applications required for the platform that are not accessible to the
end user. The MIDAS Platform runs across the health authority
network (in blue) as well as externally (in orange). Data can be
pulled into the platform from private or public data sources.
The MIDAS platform has been successfully tested, validated [6]
and evaluated by four pilot sites in public health institutes
throughout Europe (Finland, Northern Ireland, Ireland and the
Basque Region) realizing success across different priority policy areas. Each of these pilot study priorities is closely related to
COVID-19, either in terms of the associated risk (e.g. diabetes,
obesity and the ageing population), the impact of confinement
restrictions (e.g. measures of mental health and childhood
obesity), to a wide range of accessible social and spatial variables, such as measures of deprivation, social isolation, and
access to and use of healthcare services [3]. From the outset
MIDAS embedded a user-centered, co-creative, agile ap--
proach to its design and development, engaging with a
wide-range of stakeholders, having the needs of each health
policy pilot site driving the development in relation to both
data analytics and visualizations for their individually selected health policy focus [5].
Making Data Meaningful Within a Local Context

The ability to enhance the usefulness of data located within
health policy sites by integrating it with other global data

sources provides significant potential value to the user.
MIDAS provides tools such as a cross-filter analytics dashboard, an easy to use and understand interactive map visualization that updates its content automatically when the user
selects different countries on the displayed map (as shown in
Figure 3). This can be a useful tool for the initial exploration
of many types of data within a specific topic, e.g. COVID-19.
The form of visualizations is not uniform and can be customized based on the data type and research question (i.e. they
can include line charts, bar charts, maps, tables, and other
plots). The categorical variables used in the cross-filter are
predefined. Users can select subgroups for analysis and the
graphs will automatically update with data from the selected
subgroups. An example of the cross-filter tool in the MIDAS
platform is shown in Figure 3, created for the Irish pilot. The
categorical variables are Gender, Region, and Age Group, and
users can select subgroups either by clicking the buttons on
the panel, selecting the region on the map, or the subgroups
in the line chart or bar chart.
The application of the cross-filter tool can present results
more intuitively on the map and can make cross-group comparisons easier through interaction with the user. It is not necessary to have the same components as the example of the Irish
pilot shown in Figure 3, however all of the components can be
replaced according to the categorical variables and data types.
The flexibility of the tools developed for the MIDAS platform
allows the user to apply the most suitable forms of visualization
to their data.
To help users follow the results, emergency ICD-10 WHO
codes U07.1 and U07.02 from [36] have been assigned to
diagnose an identified presence of the virus. For example, the

Policy Board Users
Common User
Authentication Service
ISAACUS
Metadata
Server
GYDRA
Application
Private
Data

Public
Data
External
Network

Social Campaign Manager

Hive
Spark

MIDAS
Analytics
Application

MIDAS
Dashboard

MEDLINE Analytics
Application

HDFS
Data
Ingestion

Data Storage
and Processing

News Monitoring
Application

Data
Analysis

Data
Visualization

Health Authority Network

External
Externall Network
Networkk

FIGURE 2 MIDAS platform architecture overview, designed to host Big Data to improve decision-making in public health.

NOVEMBER 2020 | IEEE COMPUTATIONAL INTELLIGENCE MAGAZINE

53
IEEE Computational Intelligence Magazine - November 2020

Table of Contents for the Digital Edition of IEEE Computational Intelligence Magazine - November 2020
