**1. Introduction**

Extra virgin olive oil (EVOO) represents one of the most important ingredients of the Mediterranean diet, being used by most of the countries within the Mediterranean basin, owing to its excellent qualities and sensory properties ascribable to the fruits of olive trees (*Olea europaea* L.) [1]. The qualitative characteristics and the taste of EVOO are largely influenced by the olive plant varieties, the geographical origin, and the agronomic and production techniques employed as well. [2]. Recently, the consumption of EVOO has increased worldwide, even outside the Mediterranean and European countries (for example, India, Russia, China, and Australia). This trend demonstrates an increasing interest of both the producers and the consumers on the quality of food and calls for proper geographical identification and traceability of EVOOs [3].

The price of EVOO is on average 4–5 times higher than other vegetable oils. This is due to the higher production costs and to its higher nutritional and organoleptic properties. Therefore, the higher cost should in principle help to ensure high quality standards. On the other hand, the consumer is increasingly oriented towards the purchase of genuine food products with certified geographical origin [4]. In order to preserve the EVOO origin, the European Commission has established two types of certification relative to geographical origin and identification, namely the protected designation of origin (PDO) and the protected geographical indication (PGI) [4]. The definition of PGI refers to agricultural products and foodstuffs for which at least one stage of the production process must be carried out within a defined geographical area. For PDO, on the other hand, the entire production cycle must take place in a specific area. PGI labelling, therefore, focuses on quality and specific characteristics related to geographical origin [5]. As reported by the production regulations, in order to obtain the PDO certification, several conditions must be met such as a specific percentage of olive cultivars employed, well-defined cultivation practices, limited geographical areas of production, and specific characteristics regarding chemical and sensory properties of the final product. However, at the moment, to the best of our knowledge, there are no analytical parameters allowing a post hoc test on the actual geographical origin of PDO EVOOs. As a consequence, chemical and physical analyses are currently of limited use in the EVOO geographical certification [6].

Italy is one of the most important countries in the world in terms of olive oil supply and demand. Moreover, while the country boasts a large number of designations of origin labels, in total there are indeed forty-two PDOs of EVOO, the geographical indications, PGIs, exist for a limited number of three. The most active regional Italian realities are mainly located in the southern part of the peninsula, namely the region of Sicily with six PDOs and one PGI, Puglia with five PDOs, and Campania and Calabria with three PDOs each; however, there are also two important Italian central regions, such as Lazio and Tuscany, with four PDOs and one PGI, and five PDOs and one PGI, respectively [7].

Unfortunately, nowadays, one of the most common counterfeits to the detriment of the producer is the falsification of the EVOO geographical origin. Despite the great commitment and work of the authorities, counterfeiting is highly relevant for both the internal market and the global one, with difficulties in identifying and adopting convenient and reliable solutions. Indeed, due to lack of incentives, and often complex management to comply with, the technologies promoting product traceability are often difficult to implement.

Given the presence of consumers not aware of the fakes, an increasing number of low-quality olive oils often end up on the table, being not easily identifiable. The problems do not only concern consumers but also producers who operate correctly who are not economically damaged by the irregular practices of other companies [7]. Nevertheless, this type of counterfeiting causes enormous damages to the "Made in Italy" products' image and to the economy of the country.

Fake EVOO bottles often report on the label incorrect information about the product or even refer to a totally different oil.

The "Made in Italy" products are represented by a set of values enabling the consumer to distinguish them from the foreign ones. However, sometimes their advertisement is used to lead the consumer to pay an even higher price for fake qualities relative to a forged product [8]. Thus, the consumer must also be aware of the differences between fake and authentic products, being guided through tools that allow them to distinguish "what is" from "what appears to be" [4].

However, within a globalized market, the fight against these counterfeits cannot be solely based on the enhancement of consumers' awareness on the peculiarities and qualities that distinguish the "Made in Italy" from other products. Useful tools to contend with counterfeiting are those ensuring traceability [9], namely, the "possibility of reconstructing and following the path of a food in all phases of production, transformation and distribution" [10]. Therefore, traceability systems (technological and informative) are needed to strengthen and update a reliable information flow along the whole supply chain, simplifying consumers' access to information. An infotracing traceability system can integrate information related to product quality with that regarding its traceability (physical and documentary), taking advantage of an online information system [9]. As reported by Violino et al. [11], EVOO traceability is not only important to define olive oil origin, but it is fundamental for the protection against fraud. In addition, innovative tools (e.g., radio frequency identification (RFID), near field communication (NFC), and QR code technologies in combination with blockchain systems) can be commercially implemented to verify the processes and could aid in controlling the quality of virgin olive oil. An example is represented by the use of the blockchain system (a distributed database of records in the form of encrypted blocks), where the developed online information (transactions) can be protected to proof against eventual alteration and fraud [12].

Nowadays, the traceability of food products has become a priority for both consumers, who are increasingly careful to buy healthy higher quality food, and producers. Indeed, traceability can guarantee the quality of raw materials, product certification, allow a rapid identification of problematic product lots, and permit the implementation of control systems to prevent fraud. Finally, food traceability is crucial to enhance transparency for a safer internationalization of the EVOO market, with consequent fair growth of the sector [13].

In the last decade, several analytical techniques have been developed to help the identification of olive oil [14,15], and about 200 compounds, out of hundreds, have been proved to be useful as compositional markers for traceability purposes of EVOO [16]. Compositional markers include both major and minor components. State of the art EVOO traceability approaches for geographical origin assessment are represented by major components determination (e.g., triacylglycerols, triglycerides, and fatty acids), stable isotopic ratio (e.g., <sup>13</sup>C/ <sup>12</sup>C in combination with <sup>18</sup>O/ <sup>16</sup>O), and multi-element characterization through the application of different multivariate statistical techniques [3]. Those commonly used for data analysis are cluster analysis, multidimensional scaling, artificial neural networks (ANNs), and partial least squares discriminant analysis (PLSDA).

The multi-element analysis carried out by Benincasa et al. [17] allowed for a correct classification of all the organic virgin olive oils under investigation collected from different Italian regions; however, as often visible in similar studies, the method showed a high, but not excellent, percentage of correct classifications. Another example is given by the stable isotope analysis made by Portarena et al. [18], reporting an *r* ranging from 0.76 to 0.80 in distinguishing the compositions of Italian monovarietal olive oils.

Numerous analytical techniques have focused on targeted approaches for the identification and quantification of pre-defined compounds, or classes of compounds. These include gas and liquid chromatography (GC and HPLC) coupled with mass spectrometry (MS) [19,20], nuclear magnetic resonance (NMR) spectroscopy [21], infrared spectroscopy [22], fluorescence [23], inductively coupled plasma mass spectrometry (ICP-MS) [17], and DNA-based methods [24]. Conversely, limited literature is available about the assessment of olive oil adulteration using non-targeted classification approaches, focusing on the detection of all compounds in a sample without a priori knowledge of chemical entities comparable with the reference of the pure sample fingerprint profile [25,26].

Recently, fundamental research has focused on the development of non-destructive techniques to reduce the use of solvents and reagents. This is done taking into account an international context of convergence towards higher environmental sustainability and an increased human health consciousness [27]. Among various non-destructive solutions aiming to fulfill these needs, near infrared spectroscopy (NIRS) has made major achievements. NIRS, paired with chemometric techniques, were satisfactorily used for olive oil authentication and screening [28–33]. Generally, using both software and hardware, open source infostructure solutions potentially result in significant cost reduction, making the scientific tools available for a wider audience [34]. Following this path, results such as the prediction of qualitative parameters, the evaluation of indices of different fruit and vegetable products [35], the authentication of olive oil according to the variety and geographical origin [36], and the detection of adulteration through acidic composition [37] were achieved using visible/near-infrared (VIS-NIR) spectroscopy.

The aim of this work was to assess the actual geographical origin of EVOOs labeled on the market as Italian and test the potential efficiency of an open source VIS-NIR device for traceability purposes. Indeed, the device could produce results for olive oil authentication (according to its variety and origin) and for the detection of fraud in a fraction of time and potentially on a much higher sampling number with respect to conventional analytical methods. In detail, the study pursued the goal of analyzing 92 Italian and foreign EVOO samples produced in 2018 and 2019. The samples were purchased from large commercial retailers and directly from olive mills (to ensure the true origin of the product). The spectral data were analyzed with an artificial intelligence model based on neural networks.
