Fundamentals of Data Visualization

Book by: Claus O. Wilke
Source: dataviz

---

I recreated the visualization originally coded in R, from the book "Fundamentals of Data Visualization" by Claus O. Wilke, using Python and brief some important information.
Access the full book for more details.

Table of Contents:

• Fundamentals of Data Visualization
  • 1. Prerequisites
  • 2. Data
  • 3. Data visualization in Python
    • Libraries/Package
    • Tools to simplify data visualization
  • 4. Directory structure
  • 5. Issues

1. Prerequisites

To run the code, several packages stated in the requirements.sh are required. Install Miniconda3 or Anaconda before running the following command:

chmod 755 requirements.sh
./requirements.sh

After installation, a virtual environment called dataviz will be created and activated. The script requirements.sh also handled the reported issues in part 4.

2. Data

All data required for visualization practice are acquired from dviz.supp of Clause O.Wilke. For quicker load into Python, I wll covert these rda file to tsv format using the script data/rda2tsv.py and save all data in folder data/resources.

3. Data visualization in Python

Data visualization in Python like in R’s ggplot2

ggplot:

• ggplot from ŷhat
• This package has not been updated seen 2016. When installed, there raised many issues related to deprecated functions in old version of pandas.

seaborn:

• Seaborn tutorial

matplotlib:

• Matplotlib tutorial

Tools to simplify data visualization

• jpython widgets:

• Tutorial
• Blog

• holoviz: Github

4. Directory structure

.
├── 1.From_data_to_viz
│   └── 2.Mapping_data_onto_aesthetics.ipynb
├── 2.Principles_of_figure_design
├── 3.Miscellaneous_topics
├── README.md
├── data
│   └── 2_daily_temperature_NOAA.csv
├── requirements.txt
└── src
    └── utils.py

5. Issues

(1) Install rpy2 on MacOSx Using pip install rpy2 on MacOSx will turn out this error: ERROR: Failed building wheel for rpy2

Workaround: https://stackoverflow.com/a/52362473/11524628

env CC=/usr/local/Cellar/gcc/X.x.x/bin/gcc-X pip install rpy2

X.x.x is the latest version of gcc in MacOSx

(2) ggplot

All problems related to ggplot can be fixed by downgrading the version of pandas:

pip install pandas==0.19.2

To keep using the new version of pandas, the following are workarounds.

System information:

python: 3.7
pandas: 1.0.3
ggplot: 0.11.5

For current version, there is an issue when importing ggplot:

AttributeError: module 'pandas' has no attribute 'tslib'

Workaround: yhat/ggpy#662

AttributeError: 'DataFrame' object has no attribute 'sort'

Workaround: yhat/ggpy#612

AttributeError: module 'numpy' has no attribute 'ar'