The Python Ecosystem for Data Science: A Guided Tour
summary
Survey of specific libraries and workflows: NumPy pushes loops into C and yields an order-of-magnitude speedup; pandas does labeled tabular joins so a one-line pandas join can compute a Pearson correlation between population and museums in minutes; Dask mimics pandas for out-of-core scale; scikit-learn offers estimator/transformer pipelines while statsmodels supplies statistical diagnostics; NetworkX handles thousands of nodes but compiled libraries are needed for millions; Seaborn, Bokeh and Holoviews enable interactive Bokeh plots or declarative views, and JupyterHub multi-user server plus nbdime make notebook collaboration and versioning work.
AI-written summary · about
chapters (16)
- 0:00Start
- 1:24The Data Science Workflow
- 6:36Numpy
- 8:17Pandas
- 13:45Matplotlib
- 14:29Seabourn
- 15:14Iris Dataset
- 16:20Interactive Graphics
- 17:50Hollow Views
- 22:55Stats Models Library
- 24:44Network Analysis
- 26:24Started with Network Analysis in Python
- 28:49Meta Tools
- 29:28Extension to Ipython
- 33:56Jupiter Hub
- 36:26Stats Models
related talks
51mKeynote – Jake VanderPlas
33mData Science & Data Visualization in Python
46mThe Python Data Ecosystem: Navigating a fragmented landscape
43mKEYNOTE: Scaling Out PyData
45mPython for Speed, Scale, and Science
45mPython useRs
01h 31mModern scientific computing and big data analytics in Python
44mPython for Economists: An overview of Python tools for Economists