Sustainable Development Goal #4 (Quality Education) promotes “inclusive and equitable quality education” for all, including children affected by mental health, self-harm, gun violence and gang violence. Not Alone promotes the SDG Target 4.5 by supporting education “for children vulnerable situations.”
"Worldwide 10-20% of children and adolescents experience mental disorders" (World Health Organization)
Mental health and socio-economic well-being are significant to support children’s education journey and participation in an education ecosystem. Mental health illness, self-harm, gun violence and gang violence interfere with children’s ability to be integrated within an educational system.
Facebook can play a critical role in providing much-need community support and safety network for children affected by mental health, self-harm, gun violence and gang violence. Not Alone meets SDG Goal #4 by fostering a safe and healthy environment for children who are affected by mental health, self-harm, gun violence and gang violence. Not Alone helps to reduce truancy of students affected by mental illness, self-harm, gun violence and gang violence by offering proactive remedies for educators and parents, even before an incidence of mental illness, gun violence or gang violence in school.
What it does
Not Alone enhances Facebook’s AI Suicide prevention tool by expanding predictive analytics to include gun violence and gang violence in addition to suicide ideation and self-harm. Not Alone enables a community of Facebook users to detect violence before outbreak.
Not Alone provides inconstant feedback and personalized recommendations to individuals seeking help, compared to existing suicide watch chatbots (i.e. Crisis Text Line). Not Alone offers a calendarized monitoring tool to track the emotional health of Facebook users.
How we built it
We leveraged a corpus of words to conduct sentiment analysis by curating a list of positive and negative words based on scholarship on suicide prevention, gun violence and gang violence. We researched trigger word for suicide ideation, and common words that may signal gun violence and gang violence. We compiled a corpus of words to train a machine learning model on keywords that can predict incidences of suicide, gun violence, gang violence and self harm in the future.
For building predictive analytics on gang violence, we conducted a literature review and created a list of vernacular words and expressions from sociology scholarship on “Internet banging.” Internet banging refers to a culture of online behaviours by individuals perceived to be affiliated with gun violence and gang violence (Patton, Eschmann & Butler, 2013). Not Alone is significant to shed light on this under-researched area of gang-related activities on social media (Patton, Eschmann & Butler, 2013).
Challenges we ran into
We leveraged wit.ai for sentiment analysis for words. While we attempted to provide sentiment analysis for emojis, wit.ai was not able to process emojis for sentiment analysis.
Accomplishments that we're proud of
Based on our conversations with Facebook team, we broadened the relevancy of Facebook’s AI Suicide prevention tool to include gun violence and gang violence, as well as suggesting a list of proactive support provided by affected Facebook users through a chatbot.
What we learned
We learned significant potential of community safety that can be provided by Facebook’s AI Suicide prevention tool, and how the tool can be optimized to predict gun violence and gang violence.
What's next for Not Alone
We aim to increase the predictive performance of machine learning models on gun violence and gang violence that are customized for Facebook users.
World Health Organization. (n.d.). Child and adolescent mental health. Retrieved from https://www.who.int/mental_health/maternal-child/child_adolescent/en/
United Nations Educational, Scientific and Cultural Organization. (2017). Accountability in education: Meeting our commitments. Global Education Monitoring Report. Paris: UNESCO. Retrieved from https://unesdoc.unesco.org/ark:/48223/pf0000259338.page=403
Patton, D. U., Eschmann, R. D., & Butler, D. A. (2013). Internet banging: New trends in social media, gang violence, masculinity and hip hop. Computers in Human Behavior, 29(5), A54-A59.