Python as the multitool of data science
summary
Builds a data pipeline from raw JSON and CSV of the 113th Congress into a single pandas frame, then pickles it for reuse and visualizes trends with matplotlib; uses numpy for math and scikit‑learn KNN to classify yes/no votes. Pipeline processes about 316,000 rows, maps categorical fields to integers, and runs cross‑validation giving only 66% accuracy on vote prediction. Demonstrates pandas→numpy→scikit‑learn workflow and shares runnable IPython Notebook and pickled artifacts.
AI-written summary · about
chapters (17)
- 0:00Start
- 0:43Derek Bell
- 1:10Ipython Notebook
- 1:22Has Anyone Used Ipython Notebook before
- 1:38The Ipython Notebook
- 3:08Pandas
- 3:54Matplotlib
- 5:20Scikit-Learn
- 7:05Background on the Data
- 37:07Standard Deviation
- 39:27Linear Regression
- 40:46Machine Learning
- 41:48The K Nearest Neighbor Algorithm for Classification
- 45:36Python Pickling
- 45:59Cross Validation
- 47:42Scoring Function
- 50:29Does a Sk Learn Work with Pandas Dataframe
related talks
46mFrictionless Data Science w/Python
42mUsing Python pandas for scientific Research
02h 14mHands-on Python for Programmers
34mPython as a Framework for Analytics and Growth Hacking
35mWorking with data in Python for the Aspiring Data Scientists
01h 13mFrom Pandas to PySpark
01h 38mBasic Python Packages for Science
37mThe Python Ecosystem for Data Science: A Guided Tour