[PyData Global 2021] Know Your Data First: An Introduction to Exploratory Data Analysis

This is a hands-on tutorial that introduces comprehensive Exploratory Data Analysis (EDA) techniques to have better understandings about your data before doing serious tasks such as machine-learning or deep-learning.

Target Audience

• Student who wants to be a data scientist
• Junior data scientist
• Machine-learning researcher

Prerequisite

• Some experiences with
  • Python
  • Pandas
  • Matplotlib
  • Jupyter Notebook (or similar)
• GitHub & Google accounts
• Fork this repo then go to: https://colab.research.google.com/github/{your_github_id}/pydata2021-eda/

Outline

1. Introduction
2. Data loading and preprocessing
   • Loading a csv file
   • Merging many csv files
   • Essential check: #Samples, Column Names, Unique Values, Missing Values, etc.
   • sidetable
   • Preprocessing & Feture Engineering
     • Handling missing values
     • Extracting features
3. Statistical Visualizations
   • matplotlib: basic building block, essential for fine-tuning
   • pandas: data manipulation + plotting
   • seaborn: handy matplotlib wrapper for statistical visualizations
4. (Easy Enough) Interactive Visualizations
   • ipywidgets
   • plot.ly and plot.ly express
   • bokeh
   • altair
5. Automatic EDA Report
   • dtale
   • pandas-profiling
   • sweetviz
   • autoviz
6. Wrap-up and Some Tips

Contact

Sin-seok SEO @Safran Tech, Safran SA

• Webpage
• LinkedIn
• GitHub