comments from lab

**Cluster creation**
- [x] remove all inter-day correlations
- [x] remove all weird statuses (other, no response, missing)

**Assessment**
- [x] check with up to 20 clusters
- [ ] how stable are the clusters - check different seeds

**Specification**
- [x] scale using normalization
- [x] use observation window as denominator

**Descriptive statistics**
- [x] after decided specification use 2 cluster results and 5 cluster results