This is the code repository for Scalable Data Analysis in Python with Dask [Video], published by Packt. It contains all the supporting project files necessary to work through the video course from start to finish.
Data analysts, Machine Learning professionals, and data scientists often use tools such as Pandas, Scikit-Learn, and NumPy for data analysis on their personal computer. However, when they want to apply their analyses to larger datasets, these tools fail to scale beyond a single machine, and so the analyst is forced to rewrite their computation. If you work on big data and you’re using Pandas, you know you can end up waiting up to a whole minute for a simple average of a series. And that’s just for a couple of million rows! In this course, you’ll learn to scale your data analysis and execute distributed data science projects right from data ingestion to data manipulation and visualization using Dask. You’ll explore the Dask framework and see how Dask can be used with other common Python tools such as NumPy, Pandas, matplotlib, Scikit-learn, and more. You’ll be working on large datasets and performing exploratory data analysis to investigate the dataset, then come up with the findings from the dataset. You’ll learn by implementing data analysis principles using different statistical techniques in one go across different systems on the same massive datasets. Throughout the course, we’ll go over the various techniques, modules, and features that Dask has to offer. Finally, you’ll learn to use its unique offering for machine learning, using the Dask-ML package. You’ll also start using parallel processing in your data tasks on your own system without moving to the distributed environment.
- Understand the concept of Block algorithms and how Dask leverages it to load large data.
- Implement various example using Dask Arrays, Bags, and Dask Data frames for efficient parallel computing
- Combine Dask with existing Python packages such as NumPy and Pandas
- See how Dask works under the hood and the various in-built algorithms it has to offer
- Leverage the power of Dask in a distributed setting and explore its various schedulers
- Implement an end-to-end Machine Learning pipeline in a distributed setting using Dask and scikit-learn
- Use Dask Arrays, Bags, and Dask Data frames for parallel and out-of-memory computations
To fully benefit from the coverage included in this course, you will need:
This course is for data scientists, Machine Learning engineers, and data engineers who want to perform predictive analytics and data science tasks at scale. Working knowledge of Python coding and familiarity with Python libraries would be beneficial.
This course has the following software requirements: