An important element of many projects undertaken by the Dark Data Project is the derivation of actionable, structured data from analog unstructured or inferred data. To accomplish this transformation, our data scientists and analytsts leverage decades of iterative industry experience in national language processing (NLP).
This includes expertise in parsing conversational speech, slang and obfuscated lexicons, with applications for better content moderation, crisis monitoring and automated workflows.
An initiative of the Neuberger Holocaust Education Centre, Hatepedia is an online data resource used to identify and counteract the proliferation of online antisemitism and other forms of discrimination in Canada. Databases like this are critical for empowering automated natural language processing of large conversational corpi and detecting discriminatory behavior which silences marginalized voices.
YWCA Canada's #BlockHate report and polling data were commissioned to develop community-generated, survivor-centric solutions to curb the circulation of digital hate and mitigate its harms. The recommendations included in this report provide ethical guidelines and equitable benchmarks for systems-level changes to ensure online safety.
Contact us to learn more about how we're using natural language processing to help organizations better understand unstructured data.