Having recently obtained my certification as a MongoDB developer, I thought it was a good idea to do a quick write-up about my experience adopting this technology and how we can benefit from using MongoDB in machine learning projects.
The first misconception that needs the be addressed : MongoDB is not hard to pick up. When I first started learning it during a university project, my learning documentation was mainly around videos on youtube. It was not ideal due to the lack of structure. This led me to appreciate the quality of the curriculum offered by MongoDB.
The other sensitive topic is mainly targeted to companies who already have a SQL database up and running for their operations. While I agree that the cost of changing a technology stack is expensive, it does not mean it is impossible. The long term gains needs to be considered. The flexible data model can be adapted to your current structure in order to replicate it. What are the gains in that case ?
The field of AI is quickly evolving, and so are AI projects. We don't know what will be the requirements of your application to support new analytical queries. Your current data model might become obsolete quicker than you think !
Having no constraint in designing a schema enables agile development. The process of doing several iterations is adapted when we do not know where we are headed.
There are also two other features that can expend the analytical capabilities of an application. It is the topic of a next article !
To recap : I had fun doing my certification and it let me explore (pun intended) new fields in the big area of data science. Implementing MongoDB in projects is refreshing and I liked the challenge so far !