Problem 2: Social Media Flagging Solution for Potential Police Behavioral Infractions and Centralized Repository for Social Media Posts Events

[schottr1]

Theme: Police and Judicial Reform and Accountability

Brief description of your idea: Develops add-ons/extensions to social media platforms and other apps that allow the submitter or readers to flag a post as a potential police behavioral infraction and aggregate multiple posts from the same platform or different platform sources into events stored in a single repository that can be accessible via an API.

What makes your idea unique?:

• There are applications already developed where people can post pictures and information on police behavioral infractions. This would be the first cross-social media platform solution that merges multiple data sources into single Events view.
• This tool combines multiple inputs from various postings/platforms into a single view of an event.
• Using NLC/NLU, puts non-formatted data into standardized entities, concepts, and relations that can be queried and analyzed. It will use other methods to determine additional data such as location and date/time from pictures or text.
• Provides a traceback to the original source of information.
• Provides an assessment of the quality and verifiability of event information.
• Third-party applications supporting government, law enforcement, civic groups and individuals can access this data via APIs for a variety of purposes.

What would be the impact of your idea if implemented?: Problem: Victims and observers post pictures or describe incidents through non-official/governmental channels like social media platforms, or smartphone applications specifically designed for this reporting. For the same event, these postings can be from multiple people and placed on different social media platforms. In addition, these postings require individuals trying to gather information to manually find and then pull the facts from the individual postings and consolidate to provide a better picture of the event and identify if there is a likelihood of a police behavioral infraction or if this event is of a type that disproportionately targets the black community.

Benefits: By using existing social media applications and an easy method for flagging existing content, this solution will likely have greater adoption and easier/faster input than those apps requiring users to download and separately input information.

By IBM providing an API accessible, central repository of machine-curated police behavioral infraction events, government, law enforcement, and civic groups can easily integrate data from third party sources with their official systems to identify discrepancies in reporting or flag events for further investigations.

Data Scientists will be able to more easily access third-party social media sources and integrate this information with other sources to provide a more comprehensive and verifiable picture of police behavioral infractions, providing the information for not only identifying problems, but measuring the success of programs and initiatives to address disparities.

Skills to contribute (e.g. development, architecture, research, design or anything else): Development, Architecture, Design, Host solution, Leading the engagement/cooperation with social media platform companies.

[fjoenichols]

This seems like something that would not be too difficult to accomplishing leveraging existing Watson functionality. Very interesting idea.