Day 3: The Whole Game

Today we want to try going from generating a summary analysis and then taking that data into Datawrapper.

We’ll start our day with going over some of the key concepts from Orange yesterday. I’ve prepared an annotated file for you which is available here:

The dataset is available for download at this link. You will have to CMD or CTRL + S.

It is laid out like this:

Orange workflow diagram for Scooby Doo data analysis

You will:

Look in detail at what EACH STEP OF THE STAGE is doing in the pre-done, pre-answered example (red boxes).
Answer the related question by creating a new pipeline or changing values (duplicate your copy to save any progres).

Data Dictionary

Please find the data dictionary here: https://github.com/rfordatascience/tidytuesday/blob/main/data/2021/2021-07-13/readme.md

Data Rappers

We’re going to learn Datawrapper by mucking around with it first through this excellent tutorial by Lena Groeger.

Basic Data Requirements

Tabular Format: Datawrapper expects data in a tabular format, like spreadsheets or CSV files.
- Columns represent variables
- Rows represent individual entries or observations
Headers: Always include column headers in your first row.
- Use clear, descriptive names
- Avoid special characters or excessively long names
Clean Data: Datawrapper works best with pre-cleaned data.
- Remove unnecessary columns
- Fill or handle missing values appropriately
- Ensure consistent formatting across columns

Format Requirements by Chart Type

Line Charts

Need at least one column for the x-axis (usually time/dates)
Need at least one numeric column for the y-axis

Example structure:

Date,Revenue,Expenses
2023-01,5000,4000
2023-02,5500,4200
2023-03,6000,4500

Bar/Column Charts

Need one categorical column for labels
Need at least one numeric column for values

Example structure:

Country,Population
USA,331000000
China,1412600000
India,1380000000

Pie/Donut Charts

Need one categorical column for segments
Need one numeric column for values

Example structure:

Category,Sales
Product A,1200
Product B,800
Product C,600

Scatter Plots

Need at least two numeric columns (x and y coordinates)
Optional third numeric column for bubble size

Example structure:

Country,GDP,LifeExpectancy,Population
USA,65000,78.8,331000000
China,10500,77.1,1412600000
India,2000,69.7,1380000000

Appendix

Data Reshaping Cheatsheet

Melt

Purpose: Transforms wide-format data to long-format by unpivoting columns into rows.

When to use:

When you need to convert multiple columns into key-value pairs
For visualization tools that prefer long-format data
To normalize data for analysis

Example

Before (Wide Format):

   ID  Name  Math  Science
   1   John  90    85
   2   Jane  95    92

After (Long Format):

   ID  Name  Subject  Score
   1   John  Math     90
   1   John  Science  85
   2   Jane  Math     95
   2   Jane  Science  92

Pivot

Purpose: Reshapes long-format data into wide-format by pivoting rows into columns.

When to use:

When you need to create a summary table
To reshape data for reporting or visualization
To create crosstabs or contingency tables

Example

Before (Long Format):

   ID  Name  Subject  Score
   1   John  Math     90
   1   John  Science  85
   2   Jane  Math     95
   2   Jane  Science  92

After (Wide Format):

   ID  Name  Math  Science
   1   John  90    85
   2   Jane  95    92

Transpose

Purpose: Flips data so rows become columns and columns become rows.

When to use:

When you need to quickly flip the orientation of your data
For certain statistical analyses that require variables in rows
When preparing data for specific visualization formats

Example

Before:

     A  B  C
Row1  1  4  7
Row2  2  5  8
Row3  3  6  9

After:

     Row1  Row2  Row3
A    1     2     3
B    4     5     6
C    7     8     9

Quick Comparison

Operation	Input → Output	Preserves All Data	Aggregation	Common Use Case
Melt	Wide → Long	Yes	No	Converting multiple measure columns to rows
Pivot	Long → Wide	No (needs aggregation if duplicates)	Optional	Creating summary tables, crosstabs
Transpose	Flips entire table	Yes	No	Simple rotation of data structure

Visual Representation

Melt

Wide Format:          Long Format:
+----+------+----+    +----+------+--------+-------+
| ID | Name | A  | B  | ID | Name | Column | Value |
+----+------+----+    +----+------+--------+-------+
| 1  | John | 10 | 20 | 1  | John | A      | 10    |
+----+------+----+ => +----+------+--------+-------+
                      | 1  | John | B      | 20    |
                      +----+------+--------+-------+

Pivot

Long Format:            Wide Format:
+----+------+----+----+ +----+------+----+----+
| ID | Name | Col| Val| | ID | Name | A  | B  |
+----+------+----+----+ +----+------+----+----+
| 1  | John | A  | 10 | | 1  | John | 10 | 20 |
+----+------+----+----+ +----+------+----+----+
| 1  | John | B  | 20 |
+----+------+----+----+

Transpose

+-----+----+----+       +-----+-----+-----+
|     | A  | B  |       |     | Row1| Row2|
+-----+----+----+       +-----+-----+-----+
| Row1| 1  | 2  |  =>   | A   | 1   | 3   |
+-----+----+----+       +-----+-----+-----+
| Row2| 3  | 4  |       | B   | 2   | 4   |
+-----+----+----+       +-----+-----+-----+