Key questions for finding data to build ML models

Since data is the essential ingredient to power ML algorithms, it always helps to build an understanding of the patterns in data before attempting to train models. Exploratory data analysis (EDA) techniques can help build hypotheses about the data, identify data cleaning requirements, and inform the process of selecting potentially significant features. EDA can be carried out visually for intuitive insight. EDA leads naturally into feature engineering and feature selection. Feature engineer‐ ing is the process of taking raw data from the selected datasets and transforming it into “features” that better represent the underlying problem to be solved.

The constraints of data governance bring questions like:

1. What relevant datasets are available?
2. Is this data sufficiently accurate and reliable?
3. How can stakeholders get access to this data?
4. What data properties can be made by combining multiple sources of data?
5. Will this data be available in real time?
6. Is there a need to label some of the data with the “ground truth” that is to be predicted, or does unsupervised learning make sense? If so, how much will this cost in terms of time and resources?
7. What platform should be used?
8. How will data be updated once the model is deployed?
9. Will the use of the model itself reduce the representativeness of the data?
10. How will the KPIs, which were established along with the business objectives, be measured?
11. Can the selected datasets be used for this purpose?
12. What are the terms of use?

Reference: Introducing MLOps How to Scale Machine Learning in the Enterprise: Mark Treveil and the Dataiku Team