- Longtitude and Latitude as an example
- Mention deep learning
- Adding different data sets. They're often from public sources.
- Data cleansing / munging is huge. It's 80% of data science.
- Not all attributes are created equal. In fact, they are dramatically unequal. They are also only identified using ML trial and error. Design and HIPPO won't help you here.
- Goes with the idea that data is abundant.