What are the perfect steps to follow to complete Data Science Project and Machine Learning with Deployment to Cloud.
Example like -
- Checking Null values
- Outliers
Which steps comes when ?
There are no rules for steps to solve a machine learning problem (order of different steps isnt imporatant…) still u may follow the below steps
- Data collection
- Data cleaning
3)Feature Engg
4)Train test slpit
5)model fitting - prediction
- Evaluation
- Deployment
for more details go though this blog
https://towardsdatascience.com/9-steps-for-solving-data-science-problems-dc3c238cb58c
https://medium.com/cracking-the-data-science-interview/how-to-think-like-a-data-scientist-in-12-steps-157ea8ad5da8
This is a bit strange starting point for a machine learning project. Usually you start with a problem to solve, and find data that can answer that problem, etc., but, if that’s how you roll, fine.
- Given the two options you have, who are the target consumers of this machine learning model? Internal people, or external people?
- How you get the data to power your machine learning model?
- Can you use a standard vanilla algorithm or do you need to implement custom solution? Your vendor may not be able to support heavily customized solution.
- How often do you need to update the model? - rewriting the model becomes a real headache if you need to update it often.