Coding So Easy, Even a Marketer Could Do It! An Overview of Moez Ali’s In-Depth PyeCaret Tutorial
The future is written in code, that much is certain, but to effectively utilize that code without worry and stress-free is still a concerning problem for the inexperienced layman who doesn’t boast an impressive coding or data science background. The potential and value of having such skills at your fingertips is enticing, but the learning curve is often thought to be too steep.
Moez Ali, the founder and author of PyCaret admitted on his DotsLive stream that on the first day of his stats class in January of 2019, he asked the professor “What is a standard deviation?”; for he was a complete beginner in the subject. And now a year later we find the former-charter accountant a data analytics leader for PWC Canada, an avid open source contributor in his Github community, and a mentor who shares his tips and techniques both on Medium and on his extremely resourceful website, PyCaret.org; effectively proving that anyone can bridge their knowledge gap if they have the right mindset and the right process.
But Ali didn’t stop there. Having come from a non-data science background, he understood that the current channels are not business-friendly, that having domain expertise was really the only way to properly take advantage of what machine learning has to offer. This in part has to do with the fact that machine learning is an iterative process that depends on an immense amount of time and experimentation, especially if you want to improve upon said-code by an exponential degree. The machine learning life-cycle is granular in and of itself, and with so many variables (especially when you are dealing with hundreds of thousands of lines of code), a “loss of focus” can occur. Ali and his team of citizen data scientists then wondered what they could do to address this problem — how could they allow for seamless pipelines to be created, but still allow for manageable personalization and reproduction?
But after much experimentation of their own, the community launched PyCaret, An open-source, low-code machine learning library in Python that aims to reduce the cycle time from hypothesis to insights. Pycaret can be used to rapidly develop and deploy machine learning pipelines, is extremely easy to use and ultimately is a productivity tool; it is not only targeting beginners, it even is effective for data scientists. This was proven to be true when expert data scientists who were not the initial target market, saw the great value of using PyCaret for projects, for they were able to save time and still could easily customize to the degree they wanted.
Ali’s demo gave a step-by-step comparison of the same process being executed in both PyCaret and the other popular machine learning library Scikit-learn. Scikit-learn used about 150 lines of code while PyCaret needed only 23 lines to complete the exact same process with equal results. The efficiency increase is undeniable; in this tutorial, Ali walks you through several options that display the same level of impressive customization, each just as intuitive as the last. And it is only getting better, especially with the launch of version 2.0 just earlier this month.
To end off, Ali gives key advice if you want to start your PyCaret journey. The main takeaway? If you don’t have a computer science background, don’t let it discourage you! Everyone, including Ali himself, is constantly learning and improving every day. All you have to do is start small, get some hands-on experience, and properly realize that PyCaret is indeed revolutionizing coding for the future, and everyone is going absolutely crazy for it.
For a more in-depth look at PyCaret, Watch the full informative tutorial given by Ali himself here for Free on DotsLive: https://beta.dotslive.com/#/replay/06oKNVyFoFE