1. Reproducibility
1.1. Problem
1.1.1. Cause
1.1.1.1. Non - programmer
1.1.1.1.1. Bad annotation
1.1.1.1.2. Bad version control
1.1.1.2. No exact same data set with exact same format
1.1.1.3. Random seed
1.1.1.4. Limitation of size of the paper
1.1.1.5. High variability
1.1.1.6. Poor documentation
1.1.1.6.1. Delete Exploration
1.1.1.6.2. Not enough annotations
1.1.2. Effect
1.1.2.1. Re write the code
1.1.2.2. Time consuming
1.2. Benefits
1.2.1. Advancement of Science
1.2.2. Citation
1.3. What it needs
1.3.1. easy explanation
2. Who's the reader
2.1. Students(Novice)
2.1.1. More context
2.2. Colleague(Same field)
2.2.1. Freely use technical terms and explain technical details
2.2.2. Trust issue
2.3. Colleague(Different field)
2.3.1. Less technical terms and context
2.4. Business professional
2.4.1. Qualitative data only
2.4.2. Powerpoint/text report
2.4.3. Trust what data scientist does
2.4.4. Visualization
3. Length
3.1. Short and simple
3.1.1. Tend to separate to other documents if it becomes to long
3.1.2. Can't include all the details or explorations that were made
4. Data science work flow
4.1. Data cleaning 80%
5. Data Exploration
5.1. Necessary
5.1.1. Scientists who want to reproduce the research
5.1.2. Dead ends are important for future research
5.2. Unnecessary
5.2.1. Who wants to skim the work
5.3. Limitations
5.3.1. Current paper format
5.3.2. Peer reviewing system
5.3.3. Human nature to sell their work
6. New feature
6.1. Auto complete
6.2. Interactive data exploration
6.3. Drag and drop
7. Data comic
7.1. Engaging
7.2. More information in one page
7.3. Different level of importance using panels
7.4. Remember longer
7.5. Overcome limits of text
8. Documentation
8.1. Code+Markdown
8.1.1. Psychological role
8.1.2. Incentive and motivation
8.2. Poorly documented report
8.2.1. affects reproducibility
9. Jupyter notebook
9.1. Improvements
9.1.1. Cell slide
9.1.2. Interactive data exploration
9.1.3. Functions in IDE
9.2. Cons
9.2.1. Scrolling cause dizzying effects
9.2.2. not suitable for complex projects
9.3. Pros
9.3.1. Dragging cells
9.3.2. perfect for class projects, report, simple analysis
9.3.3. School, learning purpose