Skip to main content

Data and Image + R Fundamentals

Key Values of Visualisations​

  • Record information
  • Communicate information to others
  • Analyse data to support reasoning

Data Models​

  • Formal descriptions
  • Characterize data through three components
    • Objects (Item of interest)
      • Students, course, semesters
    • Attributes (properties of data)
      • Name, age, id, data, score
    • Relations (how two or more objects relate)
      • Student takes course, course during semester

N - Nominal (labels or Categories)

  • Operations =,≠=, \ne
  • e.g. maths, art(course)

O - Ordered

  • Operations =,≠,<,>=, \ne, <, >
  • e.g. A, A-, B+, B (Grade)

Q - Interval (Location of zero arbitrary)

  • Operations =,≠,<,>,−=, \ne, <, >, -
  • Can measure distances or spans
  • e.g. (3.23, -1.2) (GPS)

Q - Ration (Zero fixed)

  • Operations =,≠,<,>,−,%=, \ne, <, >, -, \%
  • Can measure ratios or portions
  • e.g. 20, 19, 22, 21 (age)

Data Processing​

  • Data cleaning and filtering
    • For quality control
    • Remove (outlier, missing data)
    • Modify (conversion of format)
  • Data adjustment
    • Depends on your task and questions to answer
    • Relational algebra
      • Aggregation, mean, sort, projection
    • Reformatting and Integration

Data Cleaning and Filtering​

  • Missing Data - no measurements, redacted
  • Erroneous Values - misspelling, outliers
  • Type Conversion - zip code to lat-lon
  • Entity Resolution - different values for the same thing
  • Data Integration - effort/errors when combining data

Dimensions and Measures​

Dimensions (Independent Variables)

  • Discrete variables describing data (N,O)
  • Categories, data, binned questions Measures (Dependent Variable)
  • Data values that can be aggregated
  • Numbers to be analysed
  • Aggregate as sum, count, avg, std. dev

Visual Language​

  • Is a sign language
  • Images perceived as a set of signs
  • Sender encodes information in signs
  • Receiver decodes information from signs

Bertins Levels of Organisation​

Data Manipulations​

  • dplyr - A grammar of data visualisation
  • Very intuitive, fast, easy for those migrating from SQL
  • When written well, code is like a recipe
  • Code the way you think

Takes the %>% operator and uses it to great effect to manipulate data frames, 5 basic 'verbs' that works with 90% of data

  • filter()
  • arrange()
  • select()
  • mutate()
  • summarise()