Why Do We Use Python In Data Science?

In two words: simplicity and versatility.

Fernando Tadao Ito
birdie.ai
2 min readFeb 17, 2021

--

Photo by David Clode on Unsplash

Python and Data Science form the perfect couple: with tons of open-sourced libraries, simple syntax, and support from a massive AI community, you can fast-track any project from inception to deployment. All industries that invest in AI may be interested in Python developers.

This language excels in scripting, making it perfect for ETL pipelines (Extract/Transform/Load), training and testing new Machine Learning models, and creating self-contained Proofs-of-Concept. These three tasks pretty much define the daily routine of a Data Scientist in any corporation, making Python the language of choice for this area.

But the best part of Python is its versatility. You can do everything running Python: from web crawling to data enrichment to back-end and even web development, you can create it all with a single language. It gives the developer a wide variety of paths for specialization and growth. A Python developer in an AI company can have any hat he desires.

A very useful tip I would give to anyone starting their studies in Python is to always keep in mind the basics of Software Engineering: code optimization, testing, and debugging.

The simplicity and permissiveness of the language may lead developers to become comfortable with its weaknesses, and bugs crawl from the cracks of the project architecture. Memory leaks, slow processes, and underutilized/bloated infrastructure are some of the most common causes of failure in Python projects, all rectifiable with proper knowledge.

Our team here at birdie.ai uses Python in almost everything that is Data: data collection uses the Scrapy engine, data pipelines are sped-up with Pyspark, data enrichment uses models trained with spaCy, PyTorch, and Transformers. Everything is glued together by AWS resources, entirely supported by external libraries: Data Scientists and Engineers share Python experience and can help each other without a language barrier.

We love to share the experiences we accrue through the agile development of our products. Stay tuned for other articles and hands-on coding tutorials!

--

--

Fernando Tadao Ito
birdie.ai

Consultant Data Scientist that also moonlights as Data Engineer