A Graphical User Interface for Environmental Statistics Rudolf - PowerPoint PPT Presentation

A Graphical User Interface for Environmental Statistics Rudolf Dutter Department of Statistics and Probability Theory Vienna University of Technology, Austria Dortmund, Germany August 13, 2008 Vienna University of Technology

Book: Applied Statistical Data Analysis DAS+R a companion of this book (Wiley, 2008). Emphasis: • Spatial Data • Strong Graphical Tools • Data from Geochemistry • Little Mathematics • Provided Software: 1 DAS+R

DAS+R (future package of R) • Basis: DAS, (all) functions of R, Tcl/Tk • Emphasis: – easy to use (e.g. Rcommander) – fast reproducibility (with small changes) of the results – interactive definition of subsets (numerically and graphi- cally) – strong interaction between statistical data and spatial information – graphical analysis 2

DAS+R 3

Open Data File 4

Read DAS/Imp (Preview) 5

Read DAS/Imp (Types of Variables) 6

Make DAS Data 7

Make Subset 8

(Pre-)Select Variables 9

View Data Set • Numerically • Graphically – Histogram – Boxplots – xy-plot – Ternary Plot – Scattermatrix (Draftman) 10

View Data Set 11

Histogram 12

Boxplots 13

Boxplots 14

Xyplot: Specifications 15

Xyplot: Specifications 16

Xyplot: Specifications 17

Xyplot 18

Worksheet 19

Worksheet Menu 20

Filled Worksheet 21

Mapping: Boxplot 22

Mapping: Boxplot 23

Mapping: Proportional Plot 24

Mapping: Simple Smoothing 25

Mapping: Simple Smoothing 26

Mapping: Kriging 27

Mapping: Kriging 28

Mapping: Kriging 29

Principal Components 30

Principal Components 31

Principal Components 32

Principal Components 33

Principal Components 34

Principal Components 35

Principal Components: Export 36

Principal Components: Mapping 37

Principal Components: Mapping 38

Background: Convert 39

Background: Convert 40

Background: Convert 41

Background: Convert 42

Result: Spatial Distrib. of Cu Cu [mg/kg] Cu [mg/kg] 0,5,25,75,95,100 ● ● ● ● ● ● ● ● ● ● Percentiles ● 98 − 4080 > 13 ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● N N ● ● ● ● 18 − 98 ● <= 13 ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● 7 − 18 ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● 5 − 7 ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● 3 − 5 ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● 0 50 100 km 0 50 100 km 43

Summary • DAS+R is user friendly • enables to construct complicated R-commands via the GUI • high repeatability (fast prototyping) • using of subsets (of variables/observations) • emphasis on graphical analysis. • Finally: Embedding of new functions should not be diffi- cult!!!! 44

Outlook To do: • Complete intended methodology • Enter all the subsets functionalities in the procedures where applicable • Remove most of the bugs • Fill all help buttons • Submit the package to CRAN. 45