G2

Lokesh S.

Senior Data Scientist

Mid-Market (51-1000 emp.)

6/7/2026

"Perfect for Prototyping and EDA, but a Headache for Version Control"

5/5

What do you like best about The Jupyter Notebook?

Rapid prototyping, exploratory data analysis (EDA), and creation of fast machine learning models to check basic business assumptions before investing the time and resources needed to develop a full data pipeline.To summarise the above, Jupyter Notebook is nothing but my canvas to explore data sets and rapid prototype. It's awesome that it's cell-based and there's immediate feedback loop. It's really very useful to be able to write out a block of code, then run it, and immediately see the black and yellow graphic below that shows an intricate plot. It helps me keep a flow state in my mind while in the midst of cleaning up messy dataset or tuning hyperparameters for a machine learning model. Other types of support include markdown — because it allows me to write our thought process, assumptions, and initial findings along with the live code, and that's a lot more user-friendly when passing prototypes along to our data engineering team. Review collected by and hosted on G2.com.

What do you dislike about The Jupyter Notebook?

This is a fantastic idea for solo work, but not so SoS in the ordinary software development process. Version control is by far the biggest pain. Taking advantage of the fact that notebooks are saved as huge JSON files with various parts of the output and metadata, change-tracking via Git just becomes pain in the ass. One of the least favorite tasks for me in Jupyter is resolving merge conflicts. In addition, it does not possess the solid syntactic support of a complete IDE such as PyCharm or VS Code—feats such as sophisticated linting, in-depth debugging or refactoring need jumping through hoops with extensions. This makes it easy to run cells out of order so as to add hidden states, and can cause the very common "it works on my machine" bug when another user tries to run the notebook in sequence. Review collected by and hosted on G2.com.

What problems is The Jupyter Notebook solving and how is that benefiting you?

As a mid-sized team, we often have to jump quickly to the end of the problem solving sprint, refute business hypotheses and stay away from building full software projects.We, as a mid-sized team, have a need to validate business hypotheses in a timely fashion before investing our resources into building full software projects. Recently, our product team wanted to identify churn trends of a new feature that was introduced into our product. I did not have to create a big pipeline, but set up my Jupyter Notebook directly to our warehouse and created a quick random forest classifier. In one afternoon I was now able to see these main churn drivers through interactive visualizations. I got to present this notebook directly to stakeholders because it's so easy to lay out that it's easily read by a non-technical person. In doing so, it can save us days of overhead because we can test and prove a concept works first in the sandbox, and then write production ready code. Review collected by and hosted on G2.com.