Analysis of the food industry in social media with Netnography and text mining methods: Twitter analysis of Torku brand

In recent years, consumers have started to use social media channels frequently for product, service, and brand evaluations. This situation has turned the emerging huge, complex, and embedded data stacks into an important resource for understanding the customer. In the development of business strategies, it is extremely important to make meaningful inferences by analyzing the data detailed. From this point of view, this study aimed to examine the posts made using the #torku tag on the Twitter platform, using netnography and text mining, by choosing the Torku brand, which is one of the top 50 companies in Turkey and a brand with a large digital data. 8208 tweets shared in Turkish on Twitter between 01.01.2011 and 01.01.2021 were accessed using the Python programming language, twint library. After all the posts were pre-processed, the BERT model was used for sentiment analysis based on natural language processing. Then, the unique words on the data, which were separated as positive and negative according to the mood, were determined and their frequencies were calculated, and word clouds were created using the Python programming language, wordcloud library. When sentiment analysis was performed on 7212 posts that did not contain the word Konyaspor, it was found that 4283 were positive and 2929 were negative. The three most recurring words in all posts; local, national, beautiful words. As a result of word cloud analysis; the three most frequently repeated words in positive posts; While the words Dosta, Recep, Anadolu are the three most frequently repeated words in negative posts; sale, pity, are official words. Common words in both positive and negative posts were also analyzed and it was seen that the three most frequently repeated words were domestic, national, and sugar words. In addition to these results, the posts with the related words were examined and it was seen that the most important factor in the consumption of the brand is that it is domestic and national and that the raw material used in the use of the products is produced by the Anatolian farmer, which creates a natural and healthy perception on the consumers. It has been observed that the use of halal products has a positive effect on consumption. When we look at the words domestic (803), national (567), Nutella (211), halal (192), Banada (97), which are the five most frequently repeated among all posts; It is seen that Torku brand users code the brand together with nationalism and halal food. When the comments containing negative words were analyzed, it was seen that consumers complained about the high prices of some products.