Data Science to measure the consumer perception on food safety Ana Ktona Etleva Muça Denada Xhaja Alba Çomo Dep.Informatis, Ela Vrenozi Faculty of Dep.Informatis, Dep.Informatics, Faculty of GFS Research, Economy and Faculty of Natural Faculty of Natural Natural Tirana Agribusiness Sciences, Tirana Sciences, Tirana Sciences, Tirana elavrenozi@yaho evadashi@ubt denada.xhaja@fs alba.como@fshn. ana.ktona@fshn. o.com .ed hn.edu.al edu.al edu.al Abstract are increasingly concerned about the content of the products they consume Recent years the issue of food safety is and they require not only quality but becoming one of the most discussed above all food safety. However, these topics not only in the circles of experts, findings were not sufficient to provide but also to albanian consumers. accurate estimates for food safety Identifying the origin of the product is perception of consumers. This is the essential to have a high quality and safe reason why a study was undertaken at system. According to FAO (WHO) country level. Recognition and "access to quality and safe food is a evaluation of this situation will serve us fundamental individual right". to understand the importance of food Guaranteeing this right is an important safety, as well as the elements that priority of the Government. In the area albanian consumers consider when of food safety, veterinary and classifying a safe food. To realize the phytosanitary issues, more than two main purpose of the work we created hundred of EU laws are adopted into two questionnaires, one for consumers national legislation and several major and one for traders / farmers. Data will steps have been taken. Despite the be collected through these semi closed updates made in legislation: Do the questionnaires in 12 counties of albanian consumers feel sure for the Albania. Analysis of data will be done food they consume? Are they informed through appropriate methods of Data that ensuring the food safety is related to Science. The choice of which method to the establishment of a traceability use will be made after comparing the system for each product? To answer this performance they have over our data question, we created a questionnaire for and the knowledge that discovers each consumers and gathered some of them after the implementation on preliminary information from consumers these data. in the Tirana district. Preliminary conclusions by analyzing data collected Keywords: food safety, food labeling, from these surveys show that consumers consumer, survey, data science. 1. Introduction Consumers seek for less expensive foods that pose no risk and believe that the food safety Food safety contains features and practices that must be addressed at the farm or during should be implemented to ensure the safety of processing. However how much information do all food products. These practices include; rules the consumers require for the food products they and standards, inspections and controls of food buy? Does the food safety affects the consumer and laboratory capacity. Food manufacturers are choice on food? How important is the origin, primarily responsible for ensuring the safety of production and distribution of food to the their food products, but also processors and consumer? Do all the agricultural products need consumers have a role in this process. Food a kind of labeling and tracing system? These are safety practices should cover the entire food some questions raised during this study and to chain, including the environment, primary which we will try to answer by gathering production, processing, distribution, preparation information from the consumers and of the product ready for consumption. traders/farmers themeselves. The main goal of EU policy on food safety is to protect the interests and health of consumers 2. Methodology and materials while guaranteeing the good functioning of one single market. To prevent food-borne diseases The main purpose of this study is to recognize are involved many sectors such as; public health, and evaluate the current situation in terms of animal health, plant and agriculture. For a better food safety to the consumers food and in this prevention effective communication, exchange context, analyze and confirm hypotheses related of information and coordination of actions is to food safety concerns. required from all the sectors. Agricultural production and food security is For this purpose a survey was conducted with considered one of the strategic sectors of both consumers and farmers/traders. It is used a Albania. qualitative methodology of collecting and This sector aims to ensure safe food of high processing information, identifying and quality based on international standards, further analyzing different variables in favor of building development of agricultural production and food safety statistics. The face to face processing capacities, and also improvement of interviewing method is used as it is considered the food chain in the country. The food sector as one of the most trusted methods of gathering aimes to meet the local demand for food, information. providing employment, reducing imports of To conduct the study were used two types of agricultural products and increasing export questionnaires. One questionnaire addressed to capacities, integration into international the consumer and the other to the trader/farmer. organizations, improving rural life, sustainable A total of 70 people were interviewed, among development and fnally why not an achievement which 50 consumers and 20 traders/farmers. of economic growth in Albania. Questions directed to the consumer are mostly Production of sufficient amounts of food while about the food they consume and the food maintaining acceptable nutritional characteristics safety. The questionnaire used for to meet the growing needs of the population traders/farmers indludes questions about their remains one of the most humane but also more clients and the product origins. Demographic challenging activities for all food manufacturers. questions are included as well on both questionaires to classify the respondets and help food labeling and treacability system is a cost understand and analyze better their responses . for the business operator therefore traders/farmers who sell food without label offer 3. Actual situation of albanian consumer it cheap and in a difficult economic situation food safety most of buyers prefer to buy at a lower price. As many local studies indicate that the prices are There are several features that a food product rising but revenues and food expenses for most must have to ensure its safety. Earlier on this of Albanian families remain the same. paper we mentioned labeling and food traceability system as main factors that provide 4. Results from the survey to the consumer information about the food safety . Food must be accompanied by a The survey investigated the food purchasing production label production, which in fact is like behavior of a representative sample in Tirana a passport for the product. It provides useful district. Two questionnaires were addressed to information about the origin of the product, different persons among which 67% were composition, microbiological composition, date consumers and 33% traders/farmers. of expiry. Traceability is the process of tracing The results indicate that most of the consumers and tracking food in all stages of production, (68%) control the food products for labeling processing and distribution [Muk14]. although all of them (100%) buy food products Traceability is defined at all levels of without labels. This indicates that although production, processing and distribution of food, labeling is considered important by the raw materials, with vegetable or animal origin, consumers, it does not affect their purchasing animals that produce food or used in food behavior. A cfause for that maybe the lack of production, including the tracing of any other information about the labeling importance substances used on food production. Local food among the consumers. An interesting fact is that business operators establish a database system to interviwed males do control the food labels record all data and ensure the procedures which (77% of them) more than females (62%). enable their identification at any time.These The age of the interviewers was set to be from databases enable the identification of any food 15 to 80 years old considering the fact that they fuenisher who has supplied the business must be decision makers on buying food for the operator with specific food. Business operators family. Interviwers were divided into four age save this information for three years and than groups to make tha sample more representative. pass it to the Ministry of Agriculture, Food and The four age groups give importance on Consumer Protection at their request. controling the food labels when buying food (>60%). Many consumers are not aware of the existence Most of adolescent consumers aged from 15 to of such system and they usually do not seek this 18 years old control the food labels (73,68%) kind of information when buing food products. the same happens with young adults aged 19 to The survey conducted showed that all the 30, 75% of them control the labels when buying interviewed consumers (100% of them) buy at food. With the older interviwers the percentage least one food product withowt labeling. A cause gets lower, (63,16 % for group age 31 to 50 and for this result may be that the Albanians 62,5% for the aged between 51 and 80 years consumers prefer buing meat, fish and old). These values show that younger consumers vegetables from local farmers which most of control the food labeling more than older ones. them have no labels for the food they sell. Also This maybe considered as an indicator that the young consumers have a higher level of uncontrolled fish and meat products gets lower. information on food labeling therefore they give From 50% of interviwers with low level of more importance to it. Another hipothesis maybe education who buy fish and meat on casual that the younger consumers have less experience shops, it decreases to 16% for the intreviwers of on buying and cooking food products so they medium education and only 7% for the need to know more about the food product they interviwers with high level of education. buy. Generally respondents present themselves as Table 2. Control of food label by education level. consumers almost always careful to Percentual values characteristics of food products as testifyies the Food Education level habit of reading information reported on labels. Answer products Mediu Hig On the other hand the purchase of products of s habits Low m h excellence – those always labeled and branded – 80,0 50,00 65,00 is yet a rare habit, at least in part, it can be Control Yes 0 food 20,0 assumed, for cost reasons. There are many produts No 50,00 35,00 0 Albanians who in their food consumtions, prove labels 100, 100,00 100,00 to be not oriented to quality but on saving, Total 00 always choosing the cheapest offer without Buy fish Yes 50,00 16,00 7,00 and meat 93,0 taking into account the source. on 50,00 84,00 No 0 It is without a doubt a quota of risen consumers imporvise 100, d shops 100,00 100,00 during these years pressing economic crisis. Total 00 Table 1. Control of food label by age. Percentual This result analysis may say that , generally, to a values higher level of education corresponds a greater attention to the food provenience, quality and Control of food label labels control. Age of Yes No Total consumers 15-18 years old 73,68 26,32 100,00 19-30 years old 75,00 25,00 100,00 31-50 years old 63,16 36,84 100,00 51-80 years old 62,50 37,50 100,00 They are numerous - especially among the respondents with higher education - those who favor the labeled and controlled food consumption: 80% with higher education control the food labels, 65% of medium education and 50% with low education level. As it is obvious regardless the education level, food labeling is a Figure 1. Relationship between level of education sensitive case for most of the consumers. and food labeling control Analogous results also for fish and meat The major part of the persons interviwed (84%) consumption: as the education level gets higher declare that spend for food products half of their the percentual of buying and consuming family monthly income, considering the fact that and 86% of them ask for the origin and most of them (78%) declare themselves as controlled and/or labeled products. families with medium economic standards this Table 3. Control of label and origin according to would leave to the conclusion that those traders. Percentual vaues consumers spend a lot for food. Considering also the traders/farmers answers, Control of label Yes No Total most of them confirm that their consumers buy and origin regulary fruits and vegetables (92%) and meat or Fruits and 20,00 80,00 100,00 vegetables animal originated products (85%). Meat, milk and This confirms the habit of albanians to eat good 86,00 14,00 100,00 sub-products and give great importance to the food they consume. On the other hand for the families with In addition of food labeling problem, traders economic difficulties, it is observed that most of themselves confirm they sell food products them choose food considering the lower price without labels, at least most of them (87,50%) more than the quality. Proof of this is the and only a small part (12,50%) do not sell blooming in the albanian cities of shops, street products without labels. This means that the vendors - especially fruit and vegetables – problem is not just informing consumers about featuring low prices and poor quality. Analysis the labels and tracebility importance to assure of the survey respondents partly confirms this frood safety but also to raise awareness among issue, most of them declare to by food –fruit and traders about the food labeling and tracing. vegetables- on local or street vendors (60% of Most of traders (81%) claim that consumers do the interviwed) and the rest (40%) buy these not react to products without labels only (19%) products on regular markets or supermarkets. of traders responded that labeling of products causes reaction. Consumers are more sensitive when it comes to Both of those observations support each other to products of animal origin (meat,milk and their explain traders behavior. As it can be presumed sub-products) , 52% of the respondets buy traders may take for granted the disinterest of products at controlled markets and consumers on food labeling so they do not supermarkets, a considerable part of them tend consider relevant food labeling. to buy those products directly from the farmers (42%) which is explained by their freshness and 5. Data Science actual and future only a few of the respondents ( 6%) buy to local contribution shops or vendors. From the data collected and analysed this survey Results gathered from traders interviwed also highlights some of the main problems that need confirm there is more sensitivity on products to be solved to guarantee the food safety. coming from animals than fruits and vegetables. The very first problem ,according to results is According to traders, 80% of the consumer does the lack of information about importance of food not require information or check for labels on labeling among the consumers and traders. A fruits and vegetables and only 20% of them do. partial solution from state organs might be to The opposite for products with animal origin, organize open events or tv spots to inform the where only 14% of the consumers/buyers don’t public and by reducing costs of product labeling show interest of the products origin and labels for the food traders or producers. Another solution is to build a secure system of generated to understand the recent trends and treacability and that the product origin and it’s situations on food safety. phases are included to the product label. Of course there are many other solutions and steps 6. Call to action and the future to be taken but here are mentioned a few. What food producing and trading companies These results show that many other surveys should do to get the most from the power of should be taken in a wide range, including a Data Science: larger sample and more topics on food  Handwritten data on food products consumption and safety. Also information from should be eliminated, use digital data websites, social networks could be veryu useful instead [Wied15]. to help understand the consumers concerns on  Invest more in I.T systems and food safety and their needs for safe food. solutions, including data analysts. This could be a Big data search. Similar to other  Demand better predictive analyses areas, the amount of “food safety related” data [Wied15]. being generated by government, industry and  Move from retrospective trouble academia is increasing rapidly [Wied15] . shooting to prospective problem  Track and Trace monitoring prevention  RFID chip technologies  Use structured and unstructured sources  Sensor-based technologies of food data (Humidity, Temperature)  Ask questions and question assumptions  More testing (targeted and untargeted)  Train data scientists that can address  Social media food related issues  Outbreak detection and genome  Enter public-private partnerships that sequencing facilitate use of Data Science in food  GIS based data safety as well as food production, Having the right data and applying the right data processing and distribution. mining techniques on them, many other models  Developing online platforms for global maybe discovered and systems generating these networking, databases for information models may be built. These models may result sharing and tools to support food safety of a great importance to the food associations management. and local government organisms helping them make the right decisions and investments on food safety. Another practical facilitiy that the use of Data The construction and functioning of early Science provide for large food companies or warning systems will inevitably require the other interested parties would be building a risk involvement of domain experts. However, management system. Such a system would with the help of data mining, we can relieve involve a process control approach with a flag those experts from many time consuming system in place to signal where a new trend on tasks, and also complement their knowledge food is developing. For example, green flags that indicate normal levels of activity would with new, interesting relations. move to red if significant additional features are Using all of these steps of Data Science is also involved or combinations of contaminants occur useful to traders/farmers and consumers as they or levels of contaminant exceed set may always refer to survey results and models concentrations. This level of activity would regulation system, the Albanian food quality ideally incorporate government and company safety guarantee system would be more laboratory or food safety information complete. Food traceability system has been management system feeds. considered as an effective means of protecting Build of a digital food treacability system the interests of consumers and improving the might be a great advantage as well, giving the food safety. If food enterprises want to have a food producers the possibility to track on real- place in the highly competitive international time their product and discover early safety market, the Food Safety Traceability System problems that might occur during production or ,which gives enterprises a competitive distribution phases [ML16]. advantage, must be established, and abilities of food-tracing and data analysis must be kept, for improving the food safety management level. Conclusions Meanwhile, based on correct data, government regulators are expected to keep the over-all In recent years, Albania has made some situation of food quality safe. achievements in the development of food Data Science is without a doubt is a great partner industry. However, food safety problems, which on this field. By keeping record of digital data, could not only impact the Albanian food export gathering information through surveys or Big trade and the sustainable economic Data techniques, analyzing, generating models development, but directly threaten the health and and building systems that use Data Mining, the safety of customers, have become more critical. quality control would definitely be more With the development of social economy and the effective, thus bringing to a safe food. improvement of technologies, the food quality control system has being progressively refined because of the implementation of the food References quality authentication, the product quality supervision and the specific products [Mot15] Motarjemi, Y., Food Safety certification referring to international practice, Management. 2015 which has also significantly improved the [Muk14] Muka, A, The issue of quality of some products. Considering the lack food security in Albania. of consumer information on food labels www.dw.de, 2014 importance, identified by using Data Science, [BLKV15] Adrie J.M. Beulens ,Yuan Li, there is much more to do on this field . A Mark R. Kramer, and Jack G.A.J. labeling system needs to be established for all van der Vorst, Possibilities for types of food products, especially for those with Applying Data Mining for Early animal provenience. Warning in Food Supply Networks, Meanwhile, according to published industry 2015. standards, the development of food standards [Wied15] Martin Wiedmann, Does Big Data and quality inspection agencies have also being offer Better Solutions for Microbial progressively refined to improve the food Food Safety and Quality?, 2015 quality. Owing to promotion and [PN13] Andrea Petróczi and Declan P. implementation of the advanced quality Naughton, Food Safety Databases: management system, enhancement of the quality A New Challenge, 2013 supervision and improvement of the laws and [ML16] Yan Ma and Yongqiu Lin, Research and Exploration of the Key Elements of Food Safety Data Analysis System Based on the Food Safety Traceability System, 2016