project

Complete transparency with zero disclosure risk

• WebApp
• Input avaiable in public repo on GitHub

Minimizing disclosure risk

• Original dataset had this granularity $9000 \times 4$
• Including individuals observations
• But the WebApp only includes $5 \times 1$
• That represents “dimensionality” reduction
• Have we lost any important information?

Optimizing access to our work

• We can now open our “safe” data to the world and encourage conversation
• Visitors to our site can test hypotheses beyond our imagination
• Science can make progress on account of our “catalyst” role

Encoding all the data in parameters

• Machine learning may do a better job here
• But for now we’ll start with our speculation about what variables matter for our key hypothesis