It was the beginning of a semester, and Luděk's team with three other classmates had four small data science projects ahead of them. Each project was based on different data: categorical, numerical, image, text. After the first week of trying to collaborate via Github and daily running into painful merge conflicts, they looked for a platform that would allow them to focus on the actual data science work instead of resolving merge conflicts.
In addition, they wanted a platform with easy onboarding and without limitations, storage and computation wise, due to the different nature of their data. In less than half an hour, their team joined Deepnote, and after four demanding projects, they can not imagine going back.
university course, group projects, predicting spread of Covid-19 in Netherlands, analysing accidents in Manchester, and more
Numpy, Pandas, scikit-learn, scikit-image, Matplotlib, Seaborn, Folium
Sometimes, you need to work independently of your colleagues, and sometimes you all want to collaborate simultaneously. As a group, they first worked individually on their implementations of a particular problem and then got together to discuss it. Deepnote let them switch between these two "modes" easily.
Deepnote is like a Swiss knife that allows you to tackle any possible problem that can arise during a data science project. Luděk's team used Deepnote with very different types of data, which also varied in size. Deepnote did not fail in any of these cases, even when training their NLP models on large corpora. In addition, if you happen to need a library not available by default, Deepnote lets you know, and then it takes two clicks to pip install it and add it to the requirements file.
Luděk's team often ran into doubts about their solution. Since other groups could have a similar issue, they posted such questions via a course forum. For this, it was useful to use cell link sharing and point their professor to the given problem with a concrete example.
Apart from its main advantages, such as collaborating in real-time, it provides you with many more tools from integrations to auto-completion of your Python code. All put together, you get to focus on the data science problem while Deepnote handles everything else.
Deepnote helped Luděk's team to successfully finish four small data science projects from beginning to end. Most importantly, they could focus on the actual problem at hand during each project rather than dealing with merge conflicts or other related issues. This was especially useful for students who wanted to quickly apply theory in practice.
Finally, it was straightforward to ask for help via Deepnote when the team was stuck or unsure: they could share the whole notebook or a particular cell with someone to help them.