IEEE Geoscience and Remote Sensing Magazine - September 2017 - 25

449

Each of the value-chain entities (user, need, product,
weights (GCW), was introduced. This scheme gives priand service) was represented as a separate table. The relaority to users involved in solving global challenges and istionships among those entities within Copernicus were
sues (extreme poverty, climate change, and so on) facing
represented in separate mapping tables. For this purpose,
humanity. Examples of such users are humanitarian relief
the primary key (PK) paradigm [43] was heavily utilized
organizations and educational and research institutes.
in the database. The mapping tables contain pairs of PKs
The shortcoming of these weights is that it is very diffifrom the entity tables being mapped. For example, to repcult to define them in an objective manner. Also, the results
resent relations between users and entities, a table named
across different weighting schemes may vary significantly;
User-Need Mapping was created that contains pairs of user
hence, sensitivity analysis was done for each of the use cases
IDs and need IDs that are the PKs corresponding to users
presented. The complete list of all users with their weights is
and needs, respectively. Figure 2(a) presents the structure
presented in Table 2.
of the database.
As the figure shows, the tables contain more data fields
DATABASE DATA AGGREGATION PROCESS
than the number of attributes presented in the "Overall Use
The data-gathering process was iterative. First, source docCase Interest Score" section. Those fields are either inforuments from FP7 and H2020 research reports, inquiring
mational or calculated depending on other fields. For exsurveys, Copernicus and GMES [44]-[48], requirements
ample, the Stakeholder and Beneficiary fields in the Users
documents, and policies were analyzed, and the data for
table are flags that show whether a given user is a stakethe entity tables were synthesized. Relationships among
holder or beneficiary. An example of a calculated field is the
those entities, represented in the mapping tables, were also
Product Score field in the Products table, which contains
filled in with data from corresponding reports. Second, afthe calculated numerical score given by (11).
ter adding new entities from new resources, the mapping
Calculated fields can depend both on internal and extables were revisited, and applicable mappings were added.
ternal attributes. For example, the field NS, calculated acThis means that the analysis not only incorporated all the
cording to (3), depends on the Normalized User Score from
information concerning EU EO users and infrastructure,
the Users table. To calculate such fields, Structured Query
but also surveyed the relationships among different projLanguage [43] was used. The queries take advantage of the
ects and services created to date. Moreover, the data and
mapping tables and choose mapped values from the correits taxonomy were then validated by a large consortium of
sponding table, which, in the case of user-need mapping, is
experts from both academia and industry.
the normalized user score of all users mapped to the same
This process identified 63 EO users expressing 38 differneed (this, in turn, means that users share the same need).
ent needs, 96 EO products across six Copernicus services,
Also noteworthy are the weighting scheme-related attri131 measurements, 472 instruments, and 342 missions asbutes in the Users table. The introduction of different weightsigned to 69 agencies having 53 ground stations. The total
ing schemes was aimed at facilitating various viewpoints in
number of relations among value-chain entities is 5,105;
ranking different users. We considered four distinct schemes.
the details are shown in Figure 3.
Within the equal weights (EW) scheme, all users are considered equal; hence, the results of the ranking are indifferRESULTS
ent to these weights. The priority weights (PW) scheme diHere, we present the chosen top-use cases emerging at the
vides users into two groups, one with high priority and one
intersection of user needs and the Copernicus service. The
with low priority. Higher priority is
given to decision makers across different areas (industry, economy, and
63 Users
38 Needs
96 Products
so on). For example, governments
are assigned higher weights, and
467
226
universities are assigned lower ones.
Users with high priority are assigned
weight 2, and users with low priority
2
60
13
are assigned weight 1. In triad weights
(TW), the paradigm for weighting is
53 Ground
69 Agencies
Six Copernicus
Stations
similar to that for the PW, but more
Services
granularity is added by presenting the
third value for the weights. The scale
623
591
in this case is between 1 and 3.
891
1,656
All these schemes target decision
342 Missions
472 Instruments
makers, leaving some users unrepre131 Measurements
sented. To address this issue, another
weighting scheme, global challenges FIGURE 3. An infographic of the data in the database.
SEPTEMBER 2017

IEEE GEOSCIENCE AND REMOTE SENSING MAGAZINE

Table of Contents for the Digital Edition of IEEE Geoscience and Remote Sensing Magazine - September 2017