Abstract
This study investigated consumer perception of food insecurity by refining data collected from social media platforms. Text mining and TF-IDF were used to extract core keywords closely related to food insecurity and analyze their meanings. In addition, time series analysis and sentiment analysis were used to examine temporal and emotional changes. The analysis results showed that keywords, such as health, stress, mental, and depression, appeared frequently, indicating that food insecurity is closely related to psychological and mental problems. In addition, consumers showed high emotional sensitivity to essential nutrients, such as vitamin D, magnesium, calcium, and omega. Furthermore, stress indices and mental and physical response indices increased simultaneously during this period, indicating that food insecurity is a factor that causes emotional and physical responses. The results of the sentiment analysis showed that negative emotions (anxiety, fear, and sadness) were higher than positive emotions, suggesting that discussions related to food insecurity have a negative emotional impact.
Keywords: food insecurity, social media, big data, South Korea
1. Introduction
Food insecurity is a fundamental problem that directly impacts human survival [1] and is considered a key indicator of structural social inequality and poverty [2]. In particular, with global crises, such as the COVID-19 pandemic, the war in Ukraine, and climate change, instability in the food supply chain is increasing, and as a result, the problems of reduced food accessibility, rising food prices, and nutritional imbalances in vulnerable groups are becoming more severe internationally [3,4,5]. Korea is no exception, and chronic or temporary food insecurity has recently emerged as a social problem, especially among single-person households, low-income families, children, and the elderly. Despite being a relatively high-income country, 4.3% of South Koreans experienced food insecurity in the first year of the COVID-19 pandemic [6,7]. Previously, food insecurity has been measured through national statistics gleaned from surveys on household and dietary trends and interviews [8,9], but these methods are hindered by the limited diversity of survey subjects, response subjectivity, and above all, the difficulties inherent in documenting trends in real-time. In addition, since the distribution, diversity, personal demographics, and dynamic characteristics of the data may not be reflected by these traditional methods, it is necessary to expand sample sizes, collection scope, and dynamic tracking to reflect consumers’ feelings about food insecurity more truthfully and objectively [10,11]. Accordingly, recently, analyzing voluntarily submitted data on the Internet, such as search terms, social media posts, and mentions in online communities, has been attracting attention as a big data-based approach to detecting social problems [12,13]. In South Korea, large online platforms, such as Naver and Daum, are used for daily information searches [14,15], and users actively use such sites to obtain information on economic difficulties, food price issues, free meals, and food support systems. Accordingly, search terms related to food insecurity carry more meaning than simple information searches and can be considered digital signals that reflect social crisis situations or the needs of vulnerable groups in real time [16]. Moreover, since food insecurity is not simply a matter of supply shortage but is also linked to psychological anxiety, it would be meaningful to be able to identify anxiety levels through data analysis on social media. By analyzing increases and decreases in the volume of search terms, related keywords, and emotional (positive and negative) trends, there is the possibility of achieving the early detection of deepening social food insecurity, regional bias, and changes in policy demands. In a study using unstructured data, Koren et al. [17] visualized the spread of regional food insecurity issues using data from Twitter (now X), and Tamasiga et al. [18] also reported that big data analysis could be used to find solutions to food insecurity and security. However, in South Korea, there has been little analysis of food insecurity based on online search terms, and systematic attempts to reflect up to date data are particularly rare.
In light of this, this study tracked the flow of search terms related to food insecurity on Naver and Daum during 2024 and empirically investigated how awareness of this topic formed and spread in domestic society through these platforms. In addition, by analyzing the correlation between major keywords, related word networks, and search patterns by period and event, it aims to provide empirically based data that can be referenced in national policies, local government welfare, and private organization support systems. This research has academic and social significance in that it explores the possibility that unstructured data can contribute to the design of food welfare policies in a digital society. It is structured as follows. The first section elucidates recent research and big data analysis trends related to current food insecurity, and the next section presents the research design, including the data used and the methodology adopted. The results section highlights the main research findings, and the conclusion outlines implications, limitations, and areas of potential future research.
2. Related Studies
In efforts to tackle the problems facing modern society, achieve secure competitiveness, and promote sustainable growth, big data research has moved from the periphery to become a core strategic field [19]. This is because decision making, which once depended on intuition or experience, is now based on quantitative and scientific evidence gleaned through big data analysis [20]. Big data refer to data that are so large in volume, high in speed, and diverse that they are difficult to collect, store, manage, and analyze using existing data analysis methods [21]. Big data are not only large but also complex, generated at high speeds, and exist in various forms [22].
Since social media networks are characterized by the 5Vs of big data (velocity, volume, value, variety, and veracity), big data analysis techniques and frameworks are widely used in their analysis [23]. Traditionally, data on users’ interests and behaviors have been collected through questionnaires, and this method is still an important tool in social science, but the emergence and popularity of social networks have made it possible to collect data on user behaviors in an unprecedented way [24]. Fan et al. [25] pointed out that the rich information provided by big data has triggered the development of new statistical and computational paradigms, enabling more sophisticated scientific discoveries. In this way, big data is also being used as a tool for solving social problems, especially in the public sector, by eliminating uncertainty and improving the quality of decision-making [26].
The following are big data studies on food insecurity, which can be considered a representative keyword in the public sector. Balashankar et al. [27] extracted text features based on news articles from 1980 to 2020 and found that the risk prediction for food insecurity can be improved compared to existing ones, and Ahn et al. [28] developed keywords to predict hunger by analyzing 53,000 news articles from 9 African countries. In addition, Eskandari et al. [29] investigated conversations related to food poverty on the social media platform Twitter during the early and later stages of the COVID-19 pandemic and found that individuals’ tweets overwhelmingly contained opinions about the increase in hunger, food poverty, and food insecurity due to the pandemic. In addition, they highlighted the rapid increase in food poverty due to hoarding, food shortages, decreased food purchasing power, and the collapse of food supplies and food systems. From a similar perspective, Martin et al. [30] analyzed Twitter in 2020 and 2021 and found that mentions of food stores, food banks, and emergency preparedness increased after the declaration of the COVID-19 pandemic. Bartelme et al. [31] focused on food poverty discourse on Twitter in Germany and presented a situation in which personal and public food insecurity experiences during the COVID-19 period were shared in real time on social media. Goetz et al. [32] also predicted emotion-based indicators, analyzing 1.2 million tweets in the United States in 2020 and proving that emotion indicators, such as fear and anger, were significantly correlated with weekly food shortage rates. Lukyamuzi et al. [33] used data mining to track tweets pertaining to food insecurity in Uganda and were able to extract relevant information and allow mitigation against it.
3. Research Methodology
3.1. Data and Summary Statistics
This paper investigated keywords related to food insecurity on social media and sought to identify changes in trends following COVID-19. The collection of data was limited to blogs and “cafes” (a kind of social media forum/group) on the South Korean-based social media platforms Naver and Daum, as it was not possible from private Facebook or Instagram accounts. Naver and Daum blogs and cafes are popular platforms in South Korea, with their members sharing news from various media outlets and actively engaging in community and forum activities. Therefore, they possess the most up-to-date content. Instagram and Facebook were excluded because many of their features do not allow private access. Therefore, it was determined that collecting keywords related to food insecurity through cafes and blogs on the two platforms would be an effective method. The data used in this study were collected from 1 January to 31 December 2024 using the keyword “food insecurity”.
3.2. Methodology
This involved examining consumer perceptions of food insecurity after COVID-19 by refining data collected from online social media platforms. The keywords for data search research were selected by domain experts, who clarified the purpose of data analysis and the relevance of the keywords. Data were collected through IMC, a big data company, and the TEXTOM analysis tool (Daegu, South Korea) was used for extraction and analysis. The keywords were categorized by frequency, and Ucinet 6.0 was used to analyze meaningful relationships between them. In addition, core keywords were derived, and their meanings were analyzed through text mining and term frequency-inverse document frequency (TF-IDF). In order to analyze the interrelationships between major keywords, a network was developed based on the co-occurrence of two keywords in the same document, and the connection strength and centrality values between them were confirmed. Clustering was performed to create word groups, and the network between groups was visualized. Finally, sentiment analysis was performed to identify the correlation between positive and negative emotions associated with food delivery-related keywords.
4. Results
4.1. Content Analysis
As a result of searching data with the keyword “food insecurity”, a total of 38,255 keywords were retrieved in 2024 (see Table 1). Since data collection based on TexTOM generally reports that 1000 words per channel are appropriate, the number of keywords collected in this study was sufficient. In addition, the results of a morphological analysis showed that there were a total of 10,745 words with a frequency of 10 or more. The narrative coding for food insecurity was clustered based on four criteria, namely food/nutrition, cause/need, sentimental/response, and connection/subject (Table 2).
Table 1.
Survey of collected data.
| Data | Channel | Section | 2024 |
|---|---|---|---|
| Food insecurity | Naver | Blog | 26,160 |
| Cafe | 7637 | ||
| Daum | Blog | 4.455 | |
| Cafe | 3 |
Table 2.
Narrative coding index.
| Categories | 2024 |
|---|---|
| Food/nutrition | 116 |
| Cause/need | 99 |
| Sentimental/response | 94 |
| Connection/subject | 70 |
| Total | 379 |
4.2. Text Mining Analysis
The frequency analysis of keywords in documents extracted using the keyword “food insecurity” (Table 3) show that “food” was the most frequent keyword, followed by “insecurity”, “health”, “intake”, “stress”, “help”, depression”, “symptom”, “good”, and “mental”, demonstrating their importance in posts related to food insecurity. In addition, the TF-IDF values of keywords, such as “eat”, “effect”, “magnesium”, “brain”, “mind”, and “vitamin A” were significantly higher than other keywords, which means that they are meaningful despite not appearing frequently. TF-IDF plays an important role in short-term trend analysis by considering both the frequency of words and the irregularity of word appearance between documents, so it is expected that these keywords will be important in the food insecurity trend in 2024. It can be assumed that this was a major factor.
Table 3.
Text mining of food insecurity.
| Rank | Word | Frequency | TF-IDF | Rank | Word | Frequency | TF-IDF |
|---|---|---|---|---|---|---|---|
| 1 | Food | 76,460 | 12,185.7 | 26 | Many | 10,491 | 20,975.36 |
| 2 | Insecurity | 72,776 | 4165.255 | 27 | Problem | 10,337 | 21,392.58 |
| 3 | Health | 65,009 | 46,365.19 | 28 | Treatment | 10,219 | 23,637.09 |
| 4 | Intake | 30,770 | 35,141.11 | 29 | Improvement | 9721 | 20,697.33 |
| 5 | Stress | 28,274 | 35,392.8 | 30 | Nerve | 9546 | 21,458.17 |
| 6 | Help | 26,470 | 32,280.54 | 31 | Sleep | 9490 | 22,932.89 |
| 7 | Depression | 22,867 | 27,802.1 | 32 | Processed food | 9449 | 19,569.05 |
| 8 | Symptom | 21,042 | 32,467.51 | 33 | Brain | 9287 | 21,736.66 |
| 9 | Good | 20,672 | 29,869.37 | 34 | Exercise | 8972 | 21,365.58 |
| 10 | Mental | 18,732 | 30,658.68 | 35 | Stability | 8796 | 19,649.24 |
| 11 | Dish | 16,644 | 29,469.45 | 36 | Reduce | 8698 | 18,866.3 |
| 12 | Body | 16,120 | 27,221.66 | 37 | Balance | 8583 | 19,564.36 |
| 13 | Eat | 15,694 | 29,457.07 | 38 | People | 8498 | 19,130.61 |
| 14 | Function | 15,404 | 26,893.31 | 39 | Vegetable | 8244 | 18,024.83 |
| 15 | Effect | 14,867 | 29,867.79 | 40 | Inclusion | 8005 | 17,812.78 |
| 16 | Important | 13,790 | 24,125.14 | 41 | Mind | 7933 | 18,692.32 |
| 17 | Influence | 13,035 | 24,156.91 | 42 | Fruit | 7734 | 17,442.91 |
| 18 | Management | 12,534 | 24,690.3 | 43 | Decrease | 7417 | 17,542.98 |
| 19 | Method | 12,224 | 23,477.33 | 44 | Need | 7266 | 16,897.16 |
| 20 | Obstacle | 11,495 | 24,423.79 | 45 | Vitamin A | 7149 | 22,085.36 |
| 21 | Maintain | 11,414 | 22,444.73 | 46 | Mood | 7125 | 17,479.93 |
| 22 | Effectiveness | 11,383 | 22,940.46 | 47 | Cause | 7059 | 18,074.13 |
| 23 | Diet | 11,311 | 22,888.57 | 48 | Omega | 6842 | 19,522.23 |
| 24 | Relax | 11,287 | 22,325.13 | 49 | Causing | 6706 | 16,489.62 |
| 25 | Magnesium | 10,507 | 29,402.93 | 50 | Prevention | 6701 | 16,809.99 |
4.3. Sentimental Network Analysis
Based on the text mining analysis results, keywords related to food insecurity were analyzed using an emotional network index. The location and role of each node were analyzed using the semantic network index, and the attributes of words with high relevance were identified. The first attribute is the degree centrality, showing the closeness of the relationship between a variable and other variables. Therefore, this index can be interpreted as a factor that directly affects consumers’ emotions. The second is betweenness centrality, which shows the mediating role that a variable plays when other variables appear. Therefore, this index can be seen as a factor that greatly depends on consumers’ perceptions of emotions. The third is closeness centrality, which indicates a variable’s synergistic emotional interactions with other variables. The fourth is page rank, which shows the page rank value of a specific variable, indicating its relative popularity compared with other variables in terms of consumer emotions. Table 4 presents the results of the semantic network analysis connecting food insecurity and emotions. Consumer sentiment on food insecurity in 2024 was examined based on the degree of centrality, betweenness centrality, closeness centrality, and page rank value.
Table 4.
Sentimental network index of food insecurity.
| Rank | Word | Degree Centrality |
Betweenness Centrality |
Closeness Centrality |
Page Rank | Group | Category |
|---|---|---|---|---|---|---|---|
| 1 | Insecurity | 209 | 26.727 | 0.005 | 1.077 | 1 | Sentimental/response |
| 2 | Fatty acid | 200 | 17.547 | 0.005 | 1.035 | 1 | Food/nutrition |
| 3 | Vitamin D | 199 | 14.912 | 0.005 | 1.030 | 1 | Food/nutrition |
| 4 | Calcium | 205 | 20.221 | 0.005 | 1.058 | 1 | Food/nutrition |
| 5 | Fiber | 198 | 13.411 | 0.005 | 1.025 | 1 | Food/nutrition |
| 6 | Iron | 196 | 15.450 | 0.005 | 1.017 | 1 | Food/nutrition |
| 7 | Theanin | 176 | 11.173 | 0.004 | 0.928 | 1 | Food/nutrition |
| 8 | Melatonin | 192 | 14.440 | 0.004 | 0.999 | 1 | Food/nutrition |
| 9 | Oriental medicine | 194 | 15.274 | 0.004 | 1.008 | 1 | Food/nutrition |
| 10 | Salmon | 198 | 14.924 | 0.005 | 1.026 | 1 | Food/nutrition |
| 11 | Tryptophan | 193 | 13.947 | 0.004 | 1.003 | 1 | Food/nutrition |
| 12 | Vitamin C | 196 | 17.604 | 0.005 | 1.018 | 1 | Food/nutrition |
| 13 | Dairy product | 200 | 16.169 | 0.005 | 1.035 | 1 | Food/nutrition |
| 14 | Walnut | 197 | 14.864 | 0.005 | 1.021 | 1 | Food/nutrition |
| 15 | Spinach | 195 | 13.656 | 0.004 | 1.012 | 1 | Food/nutrition |
| 16 | Amino acid | 199 | 20.392 | 0.005 | 1.032 | 1 | Food/nutrition |
| 17 | Neurotransmitter | 197 | 21.038 | 0.005 | 1.023 | 1 | Food/nutrition |
| 18 | Antioxidant | 201 | 15.937 | 0.005 | 1.039 | 1 | Food/nutrition |
| 19 | Green tea | 197 | 14.022 | 0.005 | 1.021 | 1 | Food/nutrition |
| 20 | Yogurt | 196 | 13.918 | 0.005 | 1.016 | 1 | Food/nutrition |
| 21 | Help | 209 | 26.727 | 0.005 | 1.077 | 1 | Sentimental/response |
| 22 | Effect | 209 | 26.727 | 0.005 | 1.077 | 1 | Sentimental/response |
| 23 | Mood | 209 | 26.727 | 0.005 | 1.077 | 1 | Sentimental/response |
| 24 | Change | 209 | 26.727 | 0.005 | 1.077 | 1 | Sentimental/response |
| 25 | Psychology | 209 | 26.727 | 0.005 | 1.077 | 1 | Sentimental/response |
| 26 | Recommend | 209 | 26.727 | 0.005 | 1.077 | 1 | Sentimental/response |
| 27 | Side effect | 209 | 26.727 | 0.005 | 1.077 | 1 | Sentimental/response |
| 28 | Danger | 209 | 26.727 | 0.005 | 1.077 | 1 | Sentimental/response |
| 29 | Take effect | 209 | 26.727 | 0.005 | 1.077 | 1 | Sentimental/response |
| 30 | Positive | 209 | 26.727 | 0.005 | 1.077 | 1 | Sentimental/response |
| 31 | Insomnia | 207 | 25.099 | 0.005 | 1.068 | 1 | Sentimental/response |
| 32 | Worry | 209 | 26.727 | 0.005 | 1.077 | 1 | Sentimental/response |
| 33 | Postscript | 209 | 26.727 | 0.005 | 1.077 | 1 | Sentimental/response |
| 34 | Tiredness | 206 | 24.362 | 0.005 | 1.063 | 1 | Sentimental/response |
| 35 | Advantage | 207 | 24.716 | 0.005 | 1.068 | 1 | Sentimental/response |
| 36 | Solution | 209 | 26.727 | 0.005 | 1.077 | 1 | Sentimental/response |
| 37 | Poor | 205 | 23.119 | 0.005 | 1.059 | 1 | Sentimental/response |
| 38 | Negation | 206 | 20.610 | 0.005 | 1.062 | 1 | Sentimental/response |
| 39 | Reference | 208 | 25.018 | 0.005 | 1.072 | 1 | Sentimental/response |
| 40 | Caution | 209 | 26.727 | 0.005 | 1.077 | 1 | Sentimental/response |
| 41 | Food | 209 | 26.727 | 0.005 | 1.077 | 2 | Food/nutrition |
| 42 | Magnesium | 207 | 24.936 | 0.005 | 1.068 | 2 | Food/nutrition |
| 43 | Processed food | 208 | 22.250 | 0.005 | 1.071 | 2 | Food/nutrition |
| 44 | Vitamin B | 209 | 26.727 | 0.005 | 1.077 | 2 | Food/nutrition |
| 45 | Caffeine | 204 | 18.882 | 0.005 | 1.053 | 2 | Food/nutrition |
| 46 | Tonic | 207 | 24.751 | 0.005 | 1.068 | 2 | Food/nutrition |
| 47 | Sugar | 203 | 17.791 | 0.005 | 1.048 | 2 | Food/nutrition |
| 48 | Protein | 207 | 24.661 | 0.005 | 1.068 | 2 | Food/nutrition |
| 49 | Vitamin | 208 | 25.657 | 0.005 | 1.072 | 2 | Food/nutrition |
| 50 | Fermentation | 206 | 19.628 | 0.005 | 1.062 | 2 | Food/nutrition |
| 51 | Microorganism | 197 | 15.635 | 0.005 | 1.021 | 2 | Food/nutrition |
| 52 | Probiotics | 199 | 16.772 | 0.005 | 1.030 | 2 | Food/nutrition |
| 53 | Fish | 205 | 18.612 | 0.005 | 1.057 | 2 | Food/nutrition |
| 54 | Serotonin | 207 | 25.590 | 0.005 | 1.068 | 2 | Food/nutrition |
| 55 | Water | 209 | 26.727 | 0.005 | 1.077 | 2 | Food/nutrition |
| 56 | Coffee | 204 | 18.876 | 0.005 | 1.053 | 2 | Food/nutrition |
| 57 | Mineral | 206 | 24.057 | 0.005 | 1.063 | 2 | Food/nutrition |
| 58 | Egg | 207 | 20.579 | 0.005 | 1.066 | 2 | Food/nutrition |
| 59 | Additive | 207 | 21.220 | 0.005 | 1.067 | 2 | Food/nutrition |
| 60 | Medical supplies | 200 | 22.748 | 0.005 | 1.037 | 2 | Food/nutrition |
| 61 | Degenerate | 207 | 21.879 | 0.005 | 1.067 | 2 | Sentimental/response |
| 62 | Wellbeing | 200 | 19.383 | 0.005 | 1.036 | 2 | Sentimental/response |
| 63 | Purchase | 206 | 20.747 | 0.005 | 1.062 | 2 | Sentimental/response |
| 64 | Inconvenience | 201 | 16.336 | 0.005 | 1.039 | 2 | Sentimental/response |
| 65 | Production | 201 | 22.779 | 0.005 | 1.041 | 2 | Sentimental/response |
| 66 | Happy | 206 | 20.148 | 0.005 | 1.062 | 2 | Sentimental/response |
| 67 | Freedom | 204 | 20.464 | 0.005 | 1.053 | 2 | Sentimental/response |
| 68 | Limit | 202 | 17.696 | 0.005 | 1.044 | 2 | Sentimental/response |
| 69 | Interruption | 202 | 17.565 | 0.005 | 1.044 | 2 | Sentimental/response |
| 70 | Promote | 205 | 23.607 | 0.005 | 1.059 | 2 | Sentimental/response |
| 71 | Practice | 194 | 13.540 | 0.004 | 1.008 | 2 | Sentimental/response |
| 72 | Experience | 156 | 8.347 | 0.004 | 0.840 | 2 | Sentimental/response |
| 73 | Investment | 168 | 9.922 | 0.004 | 0.893 | 2 | Sentimental/response |
| 74 | Sharing | 198 | 18.504 | 0.005 | 1.027 | 2 | Sentimental/response |
| 75 | Efficiency | 202 | 20.074 | 0.005 | 1.044 | 2 | Sentimental/response |
| 76 | Expectation | 162 | 7.288 | 0.004 | 0.866 | 2 | Sentimental/response |
| 77 | Hard | 202 | 23.206 | 0.005 | 1.045 | 2 | Sentimental/response |
| 78 | Popularity | 201 | 23.293 | 0.005 | 1.041 | 2 | Sentimental/response |
| 79 | Action | 184 | 14.275 | 0.004 | 0.964 | 2 | Sentimental/response |
| 80 | Emphasis | 191 | 16.610 | 0.004 | 0.995 | 2 | Sentimental/response |
| 81 | Omega | 204 | 17.602 | 0.005 | 1.053 | 3 | Food/nutrition |
| 82 | Bean | 205 | 18.900 | 0.005 | 1.057 | 3 | Food/nutrition |
| 83 | Lactobacillus | 197 | 15.713 | 0.005 | 1.021 | 3 | Food/nutrition |
| 84 | Inositol | 152 | 4.112 | 0.004 | 0.820 | 3 | Food/nutrition |
| 85 | Enzyme | 193 | 14.602 | 0.004 | 1.003 | 3 | Food/nutrition |
| 86 | Salt | 191 | 12.974 | 0.004 | 0.994 | 3 | Food/nutrition |
| 87 | Organic farming | 195 | 15.610 | 0.004 | 1.013 | 3 | Food/nutrition |
| 88 | Potassium | 185 | 11.319 | 0.004 | 0.967 | 3 | Food/nutrition |
| 89 | Conquest | 205 | 19.866 | 0.005 | 1.058 | 3 | Sentimental/response |
| 90 | Liveliness | 207 | 25.082 | 0.005 | 1.068 | 3 | Sentimental/response |
| 91 | Removal | 206 | 24.870 | 0.005 | 1.063 | 3 | Sentimental/response |
| 92 | Perfection | 202 | 19.208 | 0.005 | 1.044 | 3 | Sentimental/response |
| 93 | Comfortable | 193 | 15.814 | 0.004 | 1.004 | 3 | Sentimental/response |
| 94 | Communication | 176 | 9.570 | 0.004 | 0.927 | 3 | Sentimental/response |
In relation to emotions about food insecurity, this study confirmed that discourses were formed through keywords, such as anxiety, fatty acids, vitamin D, calcium, food, processed foods, magnesium, omega, and beans. In particular, it was noteworthy that keywords with high values in the emotional analysis results were essential vitamins and minerals, such as vitamin D, calcium, and magnesium. This result shows that South Korean consumers in 2024 were involved in a discourse on their anxieties about consuming essential vitamins and minerals that stemmed from their emotions related to food insecurity. It is believed that people tend to seek to protect themselves by eating healthy nutrients as their anxiety about food increases. This appears to reflect the trend of social obsession with immunity that has continued since the pandemic. The results of the visualization were divided into three categories, namely insecurity, food, and omega (Figure 1).
Figure 1.
Sentimental network visualization of food insecurity.
4.4. Time Series Analysis
The results of the time series analysis of the weekly document volume on food insecurity are shown in Figure 2. In 2024, an average of 1300 food insecurity keywords were created per week, showing a sharp increase at the end of the year, followed by a decrease. It can be inferred that concerns were high enough to maintain at least 1000 mentions weekly. As a result of analyzing the volume of each word related to food insecurity, the frequency of food, insecurity, and health showed a high number of mentions throughout the previous month; it peaked in July and August but showed a sharp decrease in late December.
Figure 2.
Time series analysis of food insecurity.
The results of the time series analysis are shown in Table 5 and Figure 3. A time-series analysis was conducted to identify key words related to food insecurity by frequency. They show that the stress index also increased overall during the period from May to December when food insecurity increased. This can be interpreted as showing a tendency for stress levels to increase as food insecurity increases. As a result, it can be said the former reflects the emotional and psychological consequences of the latter. In addition, the mental and body indices also increased simultaneously during this period, suggesting that prolonged food insecurity led to mental instability and physical reactions. Processed foods also showed a continuous increase in the second half of the year, implying potential alternative food consumption patterns increasing among low-income classes.
Table 5.
Time series analysis of key word frequencies.
| Date | Food | Insecurity | Health | Intake | Stress | Mental | Body | Processed Food | Vegetable | Fruit |
|---|---|---|---|---|---|---|---|---|---|---|
| 2024-01-1w | 1388 | 1255 | 899 | 493 | 451 | 202 | 226 | 104 | 119 | 119 |
| 2024-01-3w | 1475 | 1309 | 863 | 597 | 455 | 246 | 239 | 85 | 114 | 123 |
| 2024-01-5w | 1366 | 1283 | 804 | 562 | 451 | 234 | 252 | 101 | 107 | 105 |
| 2024-02-2w | 1368 | 1257 | 900 | 586 | 537 | 230 | 213 | 89 | 119 | 129 |
| 2024-02-4w | 1328 | 1290 | 1048 | 510 | 534 | 288 | 302 | 209 | 140 | 132 |
| 2024-03-2w | 1341 | 1299 | 990 | 489 | 541 | 273 | 266 | 175 | 128 | 116 |
| 2024-03-4w | 1436 | 1307 | 1058 | 546 | 511 | 226 | 279 | 177 | 121 | 116 |
| 2024-04-2w | 1346 | 1414 | 1029 | 610 | 485 | 272 | 284 | 185 | 141 | 137 |
| 2024-04-4w | 1432 | 1337 | 964 | 496 | 535 | 232 | 233 | 153 | 110 | 116 |
| 2024-05-1w | 1354 | 1300 | 930 | 583 | 555 | 267 | 295 | 147 | 124 | 118 |
| 2024-05-3w | 1510 | 1337 | 866 | 534 | 483 | 292 | 259 | 126 | 118 | 111 |
| 2024-06-1w | 1480 | 1362 | 1028 | 591 | 568 | 284 | 277 | 128 | 125 | 124 |
| 2024-06-3w | 1492 | 1363 | 970 | 571 | 560 | 243 | 236 | 126 | 144 | 142 |
| 2024-07-1w | 1345 | 1391 | 1219 | 538 | 570 | 369 | 288 | 207 | 159 | 157 |
| 2024-07-3w | 1355 | 1414 | 1576 | 559 | 533 | 495 | 372 | 276 | 214 | 180 |
| 2024-07-5w | 1257 | 1344 | 1900 | 501 | 468 | 575 | 448 | 320 | 212 | 200 |
| 2024-08-2w | 1295 | 1430 | 1901 | 544 | 511 | 527 | 459 | 301 | 236 | 207 |
| 2024-08-4w | 1469 | 1376 | 1472 | 616 | 563 | 485 | 375 | 185 | 184 | 167 |
| 2024-09-2w | 1476 | 1374 | 1425 | 566 | 483 | 483 | 391 | 175 | 163 | 163 |
| 2024-09-4w | 1482 | 1413 | 1296 | 583 | 558 | 415 | 350 | 175 | 169 | 148 |
| 2024-10-1w | 1487 | 1368 | 1408 | 624 | 471 | 481 | 307 | 163 | 164 | 146 |
| 2024-10-3w | 1563 | 1477 | 1666 | 662 | 497 | 545 | 386 | 187 | 174 | 170 |
| 2024-11-1w | 1584 | 1367 | 1473 | 597 | 551 | 423 | 309 | 157 | 169 | 150 |
| 2024-11-3w | 1579 | 1475 | 1118 | 622 | 523 | 314 | 237 | 196 | 158 | 141 |
| 2024-12-1w | 1570 | 1522 | 1423 | 699 | 657 | 404 | 331 | 238 | 194 | 166 |
| 2024-12-3w | 1709 | 1706 | 1445 | 791 | 763 | 404 | 325 | 236 | 192 | 188 |
| 2024-12-5w | 687 | 778 | 600 | 468 | 314 | 114 | 97 | 130 | 91 | 82 |
Figure 3.
Time series analysis of key word frequencies.
4.5. Sentiment Analysis
Sentiment analysis is a field of text mining technology that analyzes subjective data, such as people’s attitudes and tendencies expressed in text. It refers to natural language processing technology that analyzes subjective data, such as people’s attitudes, opinions, and tendencies expressed in text. It has the advantage of being able to automatically and quickly process polarity analysis of linguistic expressions through sentiment analysis. Words were classified using TexTOM’s emotional categories, and then frequency and emotional intensity were calculated. As a result of analyzing keywords related to food insecurity in 2024 (See Table 6), it was found that the sub-emotions (sadness, disgust, fear, pain, anger, and fright) in the negative range (57.56%) were relatively higher than the sub-emotions (feeling good, joy, and interest) in the positive range (45.4%).
Table 6.
Sentiment analysis of food insecurity.
| Frequency | Frequency Percentage | |
|---|---|---|
| Good feeling | 98,959 | 37.52512 |
| Joy | 4992 | 1.89296 |
| Interest | 7964 | 3.019938 |
| Positive total | 111,915 | 42.43802 |
| Sadness | 27,726 | 10.54021 |
| Disgust | 3978 | 1.51 |
| Fear | 87,479 | 33.17 |
| Pain | 27,892 | 10.58 |
| Anger | 1938 | 0.73 |
| Fright | 2716 | 1.03 |
| Negative total | 151,799 | 57.56198 |
| Total | 163,714 | 100.0 |
5. Discussion
This study derived words closely related to the keyword “food insecurity” based on social media big data and investigated related factors. The following conclusions were drawn from the results of text mining, time series analysis, and emotional network analysis of documents related to food insecurity during 2024. It was confirmed that this is a multidimensional social problem that goes beyond simple food accessibility issues and encompasses mental health, nutrient consumption, and emotional responses. In particular, keywords, such as health, stress, mental, and depression, appeared with high frequency, which means that food insecurity is directly related to psychological and mental problems beyond simple nutritional deficiencies. In addition, the results of TF-IDF and the emotional network analysis showed that consumers’ emotional sensitivity to essential nutrients, such as vitamin D, magnesium, and calcium omega was high. This can be interpreted as showing that social interest and anxiety about immunity influenced the discourse on food insecurity. In addition, stress and mental and physical response indices also increased simultaneously during this period, which shows that food insecurity is a factor that causes emotional and physical responses. In addition, this period saw an increase in processed food consumption, which can be seen as indicating the possible seeking out of alternative eating strategies among groups with unstable incomes. The results of the sentiment analysis also showed that negative emotions (anxiety, fear, and sadness) were higher than positive emotions, indicating that discussions related to food insecurity have a negative emotional impact.
For this research, we analyzed consumer discourse related to food insecurity from various angles, thereby providing a theoretical basis for interpreting this issue beyond the simple issue of food accessibility and in a psychological, emotional, and sociocultural context. First, the findings empirically supported the fact that food insecurity is not simply a physical deficiency but that it also acts as a trigger for mental stress and anxiety. It was empirically confirmed that food insecurity is not a mere environmental constraint but that it causes continuous emotional stress, which functions as a major variable affecting psychological and physical health. This suggests that food insecurity can be treated as a precursor or moderator of mental health in future social psychology research. Second, it was confirmed that emotionally-centered keywords for specific micronutrients, such as vitamin D, magnesium, and calcium, are deeply related to food insecurity. It theoretically showed that nutrient intake is a phenomenon motivated by social and emotional demands, not simply a physiological need. Ultimately, it suggested that consumers seek to secure emotional stability and a sense of control over health recovery through specific nutrients, and that changes in health awareness centered on nutrients are an axis of food insecurity, especially in South Korea. In addition, the strengthened social discourse centered on immunity after the COVID-19 pandemic is also reflected in the emotional response to food insecurity, suggesting that food consumption is no longer a matter of survival but an expression of identity and anxiety. Third, the centrality index derived through emotional network analysis showed that certain keywords, such as anxiety, nutrients, and processed foods, were not simply mentioned with high frequency, but were instead nodes that played a key role in the connection, mediation, and diffusion of emotional information. This can be said to have contributed to improving the theoretical validity of emotion-based semantic network analysis by overcoming the limitations of existing consumer discourse analysis that relied solely on word frequency. Fourth, consumers responded more strongly to negative emotions related to food insecurity, confirming that in situations of food insecurity, consumers tend to show emotional and preventive consumption behavior based on the fear of losing health rather than rational choices to gain health. This enabled theoretical expansion by strengthening the emotion-based risk perception model in future consumer behavior studies.
As a practical implication, this study suggests that policies connected to food production should go beyond simple supply expansion and include support for emotional health and education on nutrition to ensure their effectiveness. In particular, in the midst of a prolonged economic recession and high prices, it will be necessary to design policies that consider both psychological recovery and the stability of their dietary habits for the socially disadvantaged. In addition, since consumers are sensitive to essential nutrients, a strategy that links them to keywords, such as psychological stability, immunity, and mental health, is necessary when promoting functional foods. In addition, the close relationship with major nutrients means that consumers are not simply anxious about food insecurity but are also acting in a way that seeks to control their anxiety. This means that it is evolving from a problem of avoidance to a problem of substitution, supplementation, and reassurance, which provides important insights into policies related to food insecurity. In addition, from a communication perspective, public institutions and companies should include emotional messages that relieve anxiety about food beyond simply providing information that it is rich in nutrients. From a methodological perspective, it will be necessary to establish a monitoring system to detect and respond to social emotional flows related to food insecurity early by utilizing real-time text mining and sentiment analysis technologies based on unstructured data, such as online news, social media, and communities. In conclusion, food insecurity is a multidimensional emotional and sociocultural problem, and a single supply-oriented approach is not sufficient. As a result, it is judged that such text mining technology can become a key tool in designing food and health policies.
6. Limitation and Future Studies
This study has the following limitations and, based on these, we would like to suggest future research directions. First, the analysis was performed on unstructured text data generated from the South Korean websites Naver and Daum, and the possibility of sample bias cannot be ruled out, since data collection was conducted focusing on a specific keyword, namely “food insecurity”. In other words, documents that are relevant but not directly exposed to keywords may have been excluded from the analysis. Therefore, future studies need to ensure diversity and comprehensiveness through data collection that includes food shortages and nutritional imbalances in addition to food insecurity. In addition, various sources, such as Instagram, Facebook, and X, should be integrated to enable a more multidimensional analysis. Second, since an emotional network analysis focusing on consumer responses in South Korea was included, there may be some differences from consumer sentiment or food insecurity awareness in other cultures or countries. Accordingly, the generalizability of the research results is limited. Therefore, it will be necessary to compare and analyze various countries other than South Korea to identify differences in awareness of food insecurity, thereby deriving more universal and policy-related implications. Third, there is a temporal limitation. Since this study analyzed only one year of data from 2024, long-term trend changes or accumulated changes over several years since the pandemic may not have been sufficiently reflected. Therefore, time series analysis using at least three years of data will allow for a more precise understanding of long-term trends in consumer sentiment changes related to food insecurity. Fourth, text mining and sentiment analysis are useful for identifying correlations, but they have limitations in clearly identifying causal relationships between variables. Therefore, future research will need to connect the results of sentiment analysis with data on actual consumer behavior to empirically identify the relationship between emotions and behaviors. Finally, the study will need to conduct analyses based on demographically centered datasets to ultimately derive final results.
Author Contributions
Conceptualization, H.J., H.H.Y. and M.C.; Methodology, H.J. and M.C.; software, H.J. and M.C.; validation, H.J. and H.H.Y.; formal analysis, H.J. and M.C.; Investigation and data curation, H.J. and M.C.; writing—original draft preparation, H.J. and M.C.; writing—review and editing, H.J. and M.C. All authors have read and agreed to the published version of the manuscript.
Institutional Review Board Statement
Not applicable.
Informed Consent Statement
Not applicable.
Data Availability Statement
The original contributions presented in the study are included in the article, further inquiries can be directed to the corresponding author.
Conflicts of Interest
The authors declare no conflicts of interest.
Funding Statement
This research received no external funding.
Footnotes
Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content.
References
- 1.Odoms-Young A., Brown A.G.M., Agurs-Collins T., Glanz K. Food insecurity, neighborhood food environment, and health disparities: State of the science, research gaps and opportunities. Am. J. Clin. Nutr. 2024;119:850–861. doi: 10.1016/j.ajcnut.2023.12.019. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Berry E.M. Food insecurity, social inequity, and sustainability. World Rev. Nutr. Diet. 2020;121:95–104. doi: 10.1159/000507489. [DOI] [PubMed] [Google Scholar]
- 3.Meller F.O., Quadra M.R., Santos L.P.D., Dumith S.C., Eugenio F.D., Silva T.J.D. Association between job lost and mental health outcomes during the COVID-19 pandemic and the role of food insecurity as mediator of this relationship. Cad. Saude Publica. 2024;20:e00110523. doi: 10.1590/0102-311xen110523. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Datta B.K., Phillips S. Long COVID and the higher risk of food insecurity among participants and nonparticipants of food assistance programs in the United States. J. Acad. Nutr. Diet. 2025;125:555–566. doi: 10.1016/j.jand.2024.07.171. [DOI] [PubMed] [Google Scholar]
- 5.Otekunrin O.A. A critical assessment of the interplay of conflict, hunger, poverty, and food insecurity in Africa. Food Humanit. 2025;4:100544. doi: 10.1016/j.foohum.2025.100544. [DOI] [Google Scholar]
- 6.Choi J.H. Food insecurity is associated with dietary consumption during the COVID-19 pandemic: Data from the Korea national health and nutrition examination survey 2019–2020. Nutrients. 2023;15:772. doi: 10.3390/nu15030772. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Baek S.U., Yoon J.H. The mediating role of food insecurity in the relationship between income poverty and depressive symptoms and suicidal ideation: A nationwide study of Korean adults. Soc. Sci. Med. 2025;373:117972. doi: 10.1016/j.socscimed.2025.117972. [DOI] [PubMed] [Google Scholar]
- 8.Cai J., Parker M., Tekwe C., Bidulescu A. Food insecurity and mental health among US adults during the COVID-19 pandemic: Results from National Health Interview Survey, 2020–2021. J. Affect. Disord. 2024;356:707–714. doi: 10.1016/j.jad.2024.04.043. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Aljahdali A.A., Na M., Leung C.W. Food insecurity and health-related quality of life among a nationally representative sample of older adults: Cross-sectional analysis. BMC Geriatr. 2024;24:126. doi: 10.1186/s12877-024-04716-9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Wang L., Chen L., Li C. Research on strategies for improving green product consumption sentiment from the perspective of big data. J. Retail. Consum. Serv. 2024;79:103802. doi: 10.1016/j.jretconser.2024.103802. [DOI] [Google Scholar]
- 11.Ianni M., Masciari E., Sperli G. A survey of big data dimensions vs social networks analysis. J. Intell. Inf. Syst. 2021;57:73–100. doi: 10.1007/s10844-020-00629-2. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Azzaoui A.E.L., Singh S.K., Park J.H. SNS big data analysis framework for COVID-19 outbreak prediction in smart healthy city. Sustain. Cities Soc. 2021;71:102993. doi: 10.1016/j.scs.2021.102993. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Ann M.S. Research on public sentiment of the post-corona new normal: Through social media (SNS) big data analysis. J. Converg. Cult. Technol. 2022;8:209–215. [Google Scholar]
- 14.Jeon S.W., Ahn Y.J., Ryu G.H. A study on gamification consumer perception analysis using big data. Int. J. Adv. Cult. Technol. 2023;11:332–337. [Google Scholar]
- 15.Byun H., Jeon S.W., Yi E.S. A study on the current situation and trend analysis of the elderly healthcare applications using big data analysis. J. Korea Converg. Soc. 2022;13:313–325. [Google Scholar]
- 16.Chakraborty D., Rana N.P., Khorana S., Singu H.B., Luthra S. Big data in food: Systematic literature review and future directions. J. Comput. Inf. Syst. 2023;63:1243–1263. doi: 10.1080/08874417.2022.2132428. [DOI] [Google Scholar]
- 17.Koren O., Bagozzi B.E., Benson T.S. Food and water insecurity as causes of social unrest: Evidence from geolocated Twitter data. J. Peace Res. 2021;58:67–82. doi: 10.1177/0022343320975091. [DOI] [Google Scholar]
- 18.Tamasiga P., Ouassou E., Onyeaka H., Bakwena M., Happonen A., Molala M. Forecasting disruptions in global food value chains to tackle food insecurity: The role of AI and big data analytics—A bibliometric and scientometric analysis. J. Agric. Food Res. 2023;14:100819. doi: 10.1016/j.jafr.2023.100819. [DOI] [Google Scholar]
- 19.Fathi M., Kashani M.H., Jameii S.M., Mahdipour E. Big data analytics in weather forecasting: A systematic review. Arch. Comput. Methods Eng. 2022;29:1247–1275. doi: 10.1007/s11831-021-09616-4. [DOI] [Google Scholar]
- 20.Karatas M., Eriskin L., Deveci M., Pamucar D., Garg H. Big data for healthcare industry 4.0: Applications, challenges and future perspectives. Expert Syst. Appl. 2022;200:116912. doi: 10.1016/j.eswa.2022.116912. [DOI] [Google Scholar]
- 21.Terrazas G., Ferry N., Ratchev J.C. A cloud-based framework for shop floor big data management and elastic computing analytics. Int. J. Inf. Manag. 2019;109:204–214. doi: 10.1016/j.compind.2019.03.005. [DOI] [Google Scholar]
- 22.di Bella E., Leporatti L., Maggino F. Big data and social indicators: Actual trends and new perspectives. Soc. Indic. Res. 2018;135:869–878. doi: 10.1007/s11205-016-1495-y. [DOI] [Google Scholar]
- 23.Abkenar S.B., Kashani M.H., Mahdipour E., Jameii S.M. 2021. Big data analytics meets social media: A systematic review of techniques, open issues, and future directions. Telemat. Inform. 2021;57:101517. doi: 10.1016/j.tele.2020.101517. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Martinez-Rojas M., del Carmen Pardo-Ferreira M., Rubio-Romero J.C. Twitter as a tool for the management and analysis of emergency situations: A systematic literature review. Int. J. Inf. Manag. 2018;43:196–208. doi: 10.1016/j.ijinfomgt.2018.07.008. [DOI] [Google Scholar]
- 25.Fan J., Han F., Liu H. Challenges of big data analysis. Natl. Sci. Rev. 2014;1:293–314. doi: 10.1093/nsr/nwt032. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Wu D.C., Zhong S., Song H. Tourism and hospitality forecasting with big data: A systematic review of the literature. J. Hosp. Tour. Res. 2025;49:615–634. doi: 10.1177/10963480231223151. [DOI] [Google Scholar]
- 27.Balashankar A., Subramanian L., Fraiberger S.P. Fine-grained prediction of food insecurity using news streams. Comput. Lang. 2021:1–35. doi: 10.48550/arXiv.2111.15602. [DOI] [Google Scholar]
- 28.Ahn Y., Yan M., Lin Y.R., Wang Z. HungerGist: An interpretable predictive model for food insecurity; Proceedings of the 2023 IEEE International Conference on Big Data (BigData); Sorrento, Italy. 15–18 December 2023; pp. 1–10. [Google Scholar]
- 29.Eskandari F., Lake A.A., Butler M. COVID-19 pandemic and food poverty conversations: Social network analysis of Twitter data. Nutr. Bull. 2022;47:93–105. doi: 10.1111/nbu.12547. [DOI] [PubMed] [Google Scholar]
- 30.Martin N.M., Poirier L., Rosenblum A.J., Reznar M.M., Gittelsohn J., Barnett D.J. Enhancing artificial intelligence for Twitter-based public discourse on food security during the COVID-19 pandemic. Disaster Med. Public Health Prep. 2022:1–25. doi: 10.1017/dmp.2022.207. [DOI] [PubMed] [Google Scholar]
- 31.Bartelme T., Schonfeld M., Pfeffer J. Exploring food poverty experiences in the German Twitter-Sphere. BMC Public Health. 2024;24:1398. doi: 10.1186/s12889-024-18926-8. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32.Goetz S.J., Heaton C., Imran M., Pan Y., Tian Z., Schmidt C., Qazi U., Ofli F., Mitra P. Food insufficiency and Twitter emotions during a pandemic. Appl. Econ. Perspect. Policy. 2022;45:1189–1210. doi: 10.1002/aepp.13258. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33.Lukyamuzi A., Ngubiri J., Okori W. Tracking food insecurity from tweets using data mining techniques; Proceedings of the 2018 International Conference on Software Engineering in Africa; Gothenburg, Sweden. 27 May–3 June 2018; pp. 27–34. [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Data Availability Statement
The original contributions presented in the study are included in the article, further inquiries can be directed to the corresponding author.