Data Analysis for LEL - Week 7

Tidy data

Stefano Coretta

University of Edinburgh

Types of data

  • Tabular (or rectangular) data (like spreadsheets).

  • Audio and/or video recordings.

  • Texts or transcripts.

  • Annotation (ELAN, TextGrids, …).

  • Images.

Tabular data

  • Tabular data is made of rows and columns.

  • Prefer formats like Comma Separated Values (.csv) or Tab Separated Values (.tsv) over MS Excel files. For big data, use Parquet files.

  • If you use Excel files, keep one sheet per Excel file! (Don’t have data in multiple sheet within the same Excel file).

  • Include ONE TABLE per file. (You can transform and summarise data in R).

Tabular data: DON’T

Coding data

Use explicit coding:

  • Don’t use colours to code your data! (Software like R will discard colours and any formatting).

  • Each variable to be coded should have its own column.

  • Use clear labels:

    • ACCURACY: incorrect, correct.
      • Not 0, 1.
    • DYSLEXIC: dyslexic, non-dyslexic (or control).
      • Not 0, 1 or yes, no.
    • VOWEL: a, i, u.
    • YEAR ABROAD: year_abroad, no_abroad
      • Not yes, no.

Coding data: DON’T

Coding data: DON’T

Coding data: DO

Tidy data

Tidy data

Tidy data: DON’T

Tidy data: DO

Tidy data: DO

Tidy data: DO