Data collection and analysis tools for food security and nutrition

participant, as metadata are automatically 
26 On the other hand, see Purtova (2013, 2018), according to 
whom considering personal data as no one’s property is nothing but 
an illusion as, in practice, effective (if not legal) property right on 
personal data gets appropriated in any case by companies in the 
information industry, rather than by the individuals to whom the data 
should belong.
27 Purtova (2018) goes further, proposing that “in the age of the 
Internet of Things, datafication, advanced data analytics and data-
driven decision-making, any information relates to a person in 
the sense of European data protection law” (emphasis added) and 
therefore is subject to the practice of informed consent (Purtova, 
2018, p. 42).
gathered from the device used. For example, 
when using an online survey provider or a 
phone-based interview, the Internet Protocol 
(IP) address used by the survey participant’s 
device, or the location of her or his mobile phone 
device may be automatically passed to the data 
collection service provider. To the extent that 
such metadata may be used in combination with 
other information to identify the respondent 
(individually or as member of a specific group), 
they also must be considered personal data.
27
Second, a growing amount of personal data are 
collected by private companies in the information 
and communication industry when offering 
services such as subscriptions to cellular phone 
services, social networks, software licensing, 
etc. In agriculture, there is also the case of data 
generated by devices mounted on agricultural 
machines (tractors, harvesters, milking 
machines, etc.), which are often automatically 
sent to the machine manufacturers (justified 
by the need for information to customize or 
develop new services for farmers), but these 
may reveal elements of farmers’ activities 
which may also be considered private. The way 
consent is elicited in these cases raises doubts 
as to whether people are fully aware of what 
they are consenting to, especially when being 
presented with an all-or-nothing option to click 
either “accept” or “refuse” to enable the needed 
service (see the discussion in Purtova, 2013). 
Further, it has been argued that the way in which 
consent is requested with electronic devices may 
even lead to people releasing more personal 
information – and hence increasing the risk of 
privacy violation – than if their consent was not 
requested, in what has been termed the 
control 
paradox (Brandimarte, Acquisti and Loewenstein, 
2013). Moreover, issues of equity may arise when 
the full utilization of existing data for relevant 
public objectives is impeded by the need for 
proprietary licenses or by technological barriers 

---

90 
]
DATA COLLECTION AND ANALYSIS TOOLS FOR FOOD SECURITY AND NUTRITION
that effectively prevent some population groups 
from even accessing data that may significantly 
affect them.
Second, caution should be raised against 
commercial arrangements that are framed in 
ways that imply that the service provider has the 
right to sell data collected via surveys or acquired 
otherwise (such as through audio recording 
devices or video cameras) (Kitchin, 2014a). This 
is certainly not the case for personal data (as 
opposed to anonymous data) but the question of 
who owns personal data applies also to data that 
refer, for example, to the state of the environment 
or to the extent of nature’s exploitation. In 
this respect, the rapid evolution of new data-
generating technologies raises an entirely 
new – and largely unexplored – area of ethical 
considerations. The generation of data obtained 
by gathering, codifying and storing information 
cannot be assumed to automatically assign full 
property rights to the data generator, even when 
the information has been freely provided by 
individuals who have signed an informed consent.
The reflections above are intended to highlight 
the complexity of the aspects involved in 
designing data governance institutions, and to 
explain why this is an area of active scientific and 
philosophical inquiry, with several questions that 
remain unanswered regarding both data science 
(Blum, Hopcroft and Kannan, 2017) and its ethics 
(Floridi and Taddeo, 2016).
28
Our position is that, 
morally, 
personal data can be considered in the 
same light as blood: something that individuals 
might decide to give
, when necessary, in order 
to obtain a personal service (for example, 
when own blood is given for testing for medical 
reasons), 
but also that people should be 
encouraged to donate
, when there is a clear 
indication that its use may contribute to a greater 
good (such as saving someone else’s life). What 
should be crystal clear is that 
any resale of such 
data should be deemed immoral and even 
prosecuted as illegal
.
PRIORITY OBJECTIVES FOR 
FSN DATA-GOVERNANCE 
INITIATIVES
With the above considerations in mind, let us 
discuss some of the main priorities that effective 
data governance should tackle, with specific 
reference to FSN data.
ACHIEVING ADHERENCE TO 
GLOBAL STANDARDS AND 
HARMONIZATION OF DATA
One of the key findings of Chapter 2 in this report 
is that, though there are still a few notable gaps, 
there is already a large amount of available FSN 
data. However, these data are often fragmented 
across different public and private institutions, 
or may be collected or managed using different 
protocols, making them difficult to use it. 
Therefore, it is a priority for effective governance of 
FSN data, to strengthen international coordination 
efforts to define, promote and enforce the 
adoption of global data (and associated metadata) 
standards, including of harmonized indicators, 
which are essential for comparison and to obtain 
the full potential of data.
Within the public sectors in many countries NSOs 
play a key role in governing FSN data, and many 
of them already follow international standards. 
The United Nations Statistics Division (UNSD) has 
a long history guiding the advancement of global 
statistics. The UN (2014) Fundamental Principles 
of Official Statistics (UN Resolution 2014 A/
RES/68/261) stresses the need to harmonize 
concepts and methods, to use professional criteria 
(including scientific methods and ethics) to 
collect and use data, to develop transparent rules 
and governance mechanisms and to enhance 
coordination among statistical agencies. One of 
the key areas of work in the UNSD mandate is to 
develop harmonized statistical classifications.
29
In a survey conducted in 2020, 136 countries 
reported that they have national statistical 
legislation that complies with the Fundamental 
28 See also the entire collection of articles included in The 
ethical impact of data science¸ volume 374 of the 
Philosophical 
Transactions of the Royal Society A, available at: 
https://
royalsocietypublishing.org/toc/rsta/2016/374/2083
.
29  See 
https://unstats.un.org/unsd/classifications/
.

---

[
91
5 
INSTITUTIONS AND GOVERNANCE FOR FSN DATA COLLECTION, ANALYSIS, AND USE
Principles of Official Statistics (UNSD, 2021). 
In a similar vein, the Inter Agency and Expert 
Group on SDG indicators, created by the UN 
Statistical Commission, has spent considerable 
effort promoting the adoption by all countries of 
a harmonised set of official SDG indicators when 
reporting on progress towards SDG targets.
These efforts, however, are still largely insufficient,