In a world increasingly aware of diversity and inclusion, specialized education for individuals with Autism Spectrum Disorder (ASD) has become a critical area of focus. This research aims to demonstrate the viability of Big Data and Linear Discriminant Analysis (LDA) as tools for exploring and promoting educational strategies while effectively identifying previously overlooked themes relevant to this community. Utilizing the Twitter API and advanced data mining techniques, an exhaustive analysis was conducted on discussions across platforms such as Twitter, Instagram, and YouTube, covering the period from 2022 to 2024. This resulted in a database comprising over 2 million posts related to autism and education.

The findings, derived through Latent Dirichlet Allocation (LDA), revealed a significant intersection between autism and music. This relationship was highlighted in numerous posts, indicating how music is used as both a therapeutic and educational tool to enhance communication and social skills in individuals with ASD. Music emerged not only as a pedagogical resource but also as a means of emotional and social expression for those affected, demonstrating its inclusive potential. Additionally, significant advocacy was identified among parents of children with autism, who actively utilize social media to call for the creation and improvement of specialized classrooms in public schools. These demands underscore the necessity for adaptive learning environments that can provide specific pedagogical resources and supports essential for the educational development of children with ASD. These findings underscore the urgency of educational policies that consider specialized infrastructures and programs to ensure equitable access to quality education.

The combination of quantitative and qualitative analyses enabled the disentanglement of the dynamics surrounding viral content related to autism and allowed for the evaluation of its impact on public perception and educational policies. It was observed that, while significant events such as World Autism Awareness Day (April 2) lead to an increase in altruistically focused posts, a trend toward commercialization prevails during other times of the year. Furthermore, educational topics related to autism tend to gain traction during political campaigns but decline significantly once the elections are over.

This study highlights the crucial role of social media recommendation algorithms in the dissemination of information about autism and suggests the need for adjustments to these algorithms to promote greater visibility of educational and supportive content. In conclusion, the implementation of mixed research methodologies in social media analysis can empower educators and policymakers to develop a more inclusive and well-prepared digital citizenship, essential for addressing the challenges associated with ASD in a society that is increasingly moving toward greater understanding and acceptance.

Keywords: Big Data, LDA, autism, inclusive education, network analysis.