Scientist Meets Web Dev - How Python Became the Language of Data

Explore a keynote talk from EuroPython 2016 that delves into Python's evolution as the premier language for data science. Discover how the fusion of scientific Python and conventional software practices propelled the language to the forefront of data analysis. Learn about the historical and technical aspects that contributed to Python's success in the scientific community, challenging established tools like Matlab. Gain insights into the development of scikit-learn, its technical decisions, and how it has expanded to reach a broader audience. Examine low-level technical aspects of Python's efficiency in handling large datasets, current developments in scikit-learn and joblib, and the importance of project dynamics and documentation in software success. Follow along as the speaker demonstrates practical applications by sorting EuroPython site abstracts, explores the power of numpy arrays, and discusses the interplay between numerics, control flow, and databases in data science operations. Understand the crucial role of data and computation flow in machine learning and data mining, and see how scikit-learn simplifies these complex processes for users of all skill levels.

Intro
Data science with Python is hot
We're different
Let's do something together: sort EuroPython site EuroPyton abstracts
Why we love numpy 100 000 term frequency vs inverse doc frequency
arrays are nothing but pointers A numpy array
Array computing is fast
Array computing is limited by CPU starvation
Numerics versus control flow What if there is an if
numerics vs databases
Operations on chunks Machine learning, data mining = numerics
Operations on chunks, or algorithms on chunks Machine learning, data mining = numerics
Making the data-science magic happens
Data/computation flow is crucial
Ingredients for future data flows
The Python VM is great
Scikit-learn is easy machine learning As easy as py
Difference is richness, but requires outreach