Title: What is inside the mind of teenagers on Instagram?

Authors: Nur Aini Rakhmawati; Tetha Valianta; Irmasari Hafidz; Azmi Pratama; Doddy Ridwandono; Lolanda Annisa

Addresses: Department of Information Systems, Institut Teknologi Sepuluh Nopember Surabaya, Indonesia ' Department of Information Systems, Institut Teknologi Sepuluh Nopember Surabaya, Indonesia ' Department of Information Systems, Institut Teknologi Sepuluh Nopember Surabaya, Indonesia ' Department of Information Systems, Institut Teknologi Sepuluh Nopember Surabaya, Indonesia ' Department of Information Systems, Institut Teknologi Sepuluh Nopember Surabaya, Indonesia ' Department of Information Systems, Institut Teknologi Sepuluh Nopember Surabaya, Indonesia

Abstract: Instagram is one of the most used social media with 45 million active users on each month. In Indonesia, the highest number of internet users is between the age of 13-34 years old. It shows that most of the new internet users are young people. In this research, we create a topic model for the Instagram caption of teenager users in Surabaya, Indonesia, using latent dirichlet allocation (LDA) method. By using pen and paper questionnaire, we collected total number of Instagram 494 valid accounts with 4,664 captions. The data were collected from January 2014 to June 2017. The process of modelling using LDA was performed by experimenting with a set of number of topics: 2, 3, 4 and 5. The two topics were selected because it has a small value of perplexity, which indicates that the model has a good level of conformity. The two topics represents two categories: 'school' and 'relationship'. It was found that the topic model was dominated by the 'relationship' category.

Keywords: Instagram; captions; topic model; teenager; latent dirichlet allocation; LDA.

DOI: 10.1504/IJBIS.2021.115364

International Journal of Business Information Systems, 2021 Vol.37 No.2, pp.224 - 235

Received: 10 Aug 2018
Accepted: 25 Dec 2018

Published online: 26 May 2021 *

Full-text access for editors Access for subscribers Purchase this article Comment on this article