I often find myself using the Diabetes.jmp data table for demos. It's rich in lots of features. Multicollinear predictors, resonates with the health care (providers, insurers, researchers, clinical/nonclinical settings) community, happenstance data, a disease and predictor set that is readily recognizable and familiar to people, etc. Over the years I've been compelled to modify it for a specific demo's requirements. Two requirements that are common are geographic mapping and a categorical response in addition to the classic disease progression continuous variable. So I thought I'd post a modified version here for people's use. The major modifications are:
1. I added a county variable...you can replace my counties as you see fit. Used for geographic mapping.
2. A categorical response; Retinopathy (a common affliction for those as diabetes worsens). The categorical response is strongly correlated with the disease progression response. As disease progression gets higher (clinically worsens) the probability of developing retinopathy increases.
3. I deleted all the legacy scripts from the original Diabetes.jmp data table as well. Just to clean up the data table and none of them call the Retinopathy response...so I just decided to delete them.
