Tidy up your unstructured data, and conduct exploratory data analysis. Avoid leakages and minimize noises so that your data is prepared in a way that is in line with your business objectives.
Answer your business questions with in-depth data analysis and concise but yet informative charts.
Fancy STOA models are not panaceas. You need a plan to find the model that best fits your specific objectives and constraint. A well-oiled experiment-to-production pipeline is also critical to your success.
Machine learning systems carry the biases its training data contains (computers can also be biased, too!), and they have real impacts on people's lives. We should also be aware of the ways our system can be misused and exploited.
