A platform for human-machine interaction

Social@risk™ is an intelligent, transparent system where experts and stakeholders inform the algorithms that facilitate data structuring and analysis. The tool collects grievances and information about rights violations from social media. Experts transform these voices into insights for preventive action and organising remedy. Social@risk™ delivers intelligence about factory sites, suppliers, sectors, and regions.

Open source intelligence

Rights violations and inhumane working conditions are not private matters that corporations may handle internally. Workers have the right to know. We, therefore, want the results of our projects to become open access. Open source intelligence ensures transparency, enhances accountability, and empowers workers. Yet to ensure privacy and data integrity we encrypt and anonymise raw data.

Social@risk™ consists of two core systems

Social@risk™ is a platform that develops with each project. The heart of our system consists of two elements: Social@riskBase is the core of our insight accumulation process. Dedicated crawlers screen the Internet and channel collected data into Social@riskBase. Experts read and analyse data with Social@riskLink; the second core of our system. Social@riskLink is a toolbox which contains different text structuring and clustering views.

The millions of posts in Social@riskBase provide micro-bits of information about rights violations and social risks. We program algorithms that connect the dots in order to generate topics. We can continuously update the underlying pool of posts in order to monitor topics over time.

Currently we store information in Chinese, Vietnamese, Thai, Khmer and Burmese. More languages will be added. Social@riskBase contains information on the textile industry, electronics manufacturing and the food sector. Each project broadens the scope of our data pool.

In order to ensure data security and integrity this database is encrypted and only experts can access it. Work tasks and fields define an expert’s access rights, i.e. experts can only see the data that is relevant for their particular assignment.

Experts can access the data through three different analytical lenses:

 

1.Word index view

Experts use advanced keyword filtering and trace topics with timeline graphs of posts and visuals for word frequency rankings.

 

2. Post item view

Experts read posts that are ranked with a ‘relevance algorithm’. We improve and fine-tune the algorithm with a machine learning approach.

 

3. Graph based view (experimental)

Posts are exported into a network database where each post can stand in various relations with other texts or images

Four steps towards better protection of workers’ rights

Social@risk™ is a platform for collecting, structuring, and analysing workers’ voices. We organise this process along four steps from data collection, to leverage. Our approach to data analysis and insight management establishes multiple feedback mechanisms between experts, partners, and intelligent machines. The outcome is an effective, and goal-oriented work flow.

Development of Social@risk™