UPA Package 4, Module 2

1
UPA Package 4, Module 2
DESCRIPTIVE STATISTICS AND CLASSIFICION OF DATA
2
Descriptive Statistics and Classification of Data

• Introduction statistics
• Level of measurement
• Descriptive statistics
• Classification of data
• SPSS and EXCEL
• Introduction exercises 4.2.1a/b

3
Introduction Statistics
4
Introduction Statistics

• Concepts
• Collecting (sampling), analyzing and conclusions
  about data
• Population - great variability of data
• Probability and uncertainty
• Assumptions and models
• Purpose
• On the basis of a sample draw sound conclusions
  about whole population
• Functions
• Descriptive summarizing data to make data more
  useful
• Inductive prediction, testing relationships,
  generalization on basis of sampling

5
Level of Measurement

• Measurement
• Ability to assign numbers to things according to
  a clear and well defined rule
• Are we dealing with the right things?
• Operational definition e.g. how to measure
  poverty?
• Data types
• Nominal classification of data into categories or
  classes (land use)
• Ordinal order of categories (low/middle/high
  income), no magnitude
• Interval/Ratio unit of measurement (income in /
  month)
• Know your variables !

6
Level of Measurement

• Data Transformations
• Change high level of measurement (interval) to
  lower level of measurement (ordinal or nominal).
  Advantages and disadvantages

7
Descriptive Statistics

• Summarizing Data
• Readability (limited classes) vs. accuracy (many
  classes)
• Summarizing Nominal / Ordinal Data
• Proportions
• Percentages
• Summarizing Interval Data
• Location mean, mode, median, freq. distribution
• Variation standard deviation, coefficient of
  variation

8
Descriptive Statistics
Proportions, counting (f/N) Percentages, relative
size. Proportions 100 Rate fraction,
proportion or percent
9
Descriptive Statistics

• Location
• Central tendency, typicality, degree of
  homogeneity numbers that represent the center
  or typical value of a frequency distribution
• Mode most frequent score, highest point of the
  curve
• Median middle case when arranged in order of
  size
• appropriate for skewed distributions
• Mean central tendency of the group
• extreme values disturbing effect
• Meanmodemedian in case of symmetric
  distributions

10
Descriptive Statistics

• Frequency distribution table that shows the
  frequency of observations in each category of a
  variable
• Categories should have no overlap (exclusive) and
  complete coverage
• Give examples of frequency distribution for
  nominal, ordinal and interval type of data
• Discrete and continuous data
• Class limits

11
Descriptive Statistics

• Variation
• Dispersion, degree of heterogeneity numbers
  that depict the amount of spread or variability
  in a data set
• Range difference between highest and lowest
  score
• Standard deviation square root of the mean of
  the squared
• deviations from the mean
• Coefficient of variation standard deviation (s)
  relative to the mean (x)
• cv s / x
• Variance s²

12
Descriptive StatisticsDecision Tree for the
correct use of descriptive statistics
13
Organization and Classification of Data

• Why classification
• Simplicity, clarity, particularity, data source,
  data processing
• Classification principles
• Purpose, exclusivity, exhaustiveness, detail vs.
  readability
• Number of classes and class limits (interval
  data)
• First look at distribution, avoid e.g. empty
  classes

14
Classification of Data
15
Frequency Distribution

• Frequencies (graphs, tables) first look at data
  distribution
• Patterns and deviations from that pattern,
  outliers
• Absolute, relative and cumulative freq.
  Distributions
• Categorical and quantitative variables
• histograms, bar charts, ogives, pie charts
• smooth (computerized) curves versus histograms

16
Frequency Distribution

• Frequency table
• Patterns and deviations from that pattern,
  outliers
• Absolute, relative and cumulative freq.
  Distributions

17
Frequency Distribution

• histograms, bar charts, ogives, pie charts
• smooth (computerized) curves versus histograms

18
SPSS and Excel

• Statistical Package for the Social Sciences
  (SPSS) www.SPSS.com
• Comprehensive system for analyzing data
• Investment and upgrades, modular structure,
  simple spreadsheet structure
• Spreadsheet (e.g. Microsoft EXCEL)
• Limited statistical functionality (but how much
  do you really need?)
• Relatively cheap

19
SPSS and Excel

• Explore Central Bureau of Statistics (CBS) data
  from 1999
• www.cbs.nl
• Spatial units
• Municipality, Districts, Neighborhoods
• Municipality of Enschede
• Selection and basic descriptive analysis of
  variables
• Exercise 4.2.1a use SPSS
• Exercise 4.2.1b use EXCEL