As a data scientist, I need automation of the analytics process so that I am free from:
1. Time consuming data cleaning
2. Time consuming data exploration and feature engineering
3. Ad-hoc try-and-error model selection and tuning
I also want this step automatic:
1. finding external data source that are relevant for my problem.
So I suggest:
1. link DSX to external data sources like weather data, the weather company should like this idea.
2. Automatic feature engineering: look at my own work on automation of feature engineering from relational database
3. link to Watson API for NLP capability, some users may want to use them for extracting sentiments from texts etc.