Charts and Graphs - PowerPoint PPT Presentation

1 / 16
About This Presentation
Title:

Charts and Graphs

Description:

... sales vs. temperature, hospital admissions from suburb vs. average ... year, plant growth for each month, hospital admissions from each day of the week. ... – PowerPoint PPT presentation

Number of Views:37
Avg rating:3.0/5.0
Slides: 17
Provided by: davidco151
Category:
Tags: charts | graphs

less

Transcript and Presenter's Notes

Title: Charts and Graphs


1
Charts and Graphs
  • Dr. David Cornforth
  • School of Information Technology and Electrical
    Engineering
  • University of New South Wales _at_ ADFA

2
Why use graphs?
  • To see data more clearly

3
Types of charts
4
Axes
  • The vertical axis is the dependent variable
  • The horizontal axis is the independent variable
  • We say a graph of Y against X meaning Y is
    dependent on X.
  • Mixing them up completely changes the results of
    regression.

5
Interpreting graphs what to look for
  • Are all the data points similar?
  • Does any one stand out (an outlier)?
  • Are there any obvious groups (clusters)?
  • Is there any obvious trend, or just random
    variation? How can you tell? (Correlation)
  • What applications do these graphs have?

6
Non-linear relationships
7
Non-linear relationships
8
Transforming data
  • Makes non-linear plots into straight lines
  • Helps to identify the kind of relationship

9
Correlation
  • A measure of how things are related, or connected
  • Applies to XY scatter charts
  • In Excel, use three routes
  • Tools Data Analysis Correlation
  • Click on a chart, select Chart Add Trendline
    Options Display R-squared value
  • Insert Function CORREL(array, array)
  • Get a number between -1 and 1

10
Correlation r
  • Degree of similarity between matched data
  • Useful indicator of relationships
  • Never confuse correlation with causation

11
Measures of Correlation
  • Pearson's product-moment correlation
  • Spearman's rank correlation coefficient

12
Correlation versus causation
  • Variables may be correlated for any one of
    several reasons
  • Chance
  • Structural similarity
  • Indirect causation
  • Direct causation

13
A Bad Example
  • Whats wrong with this graph?

14
Abusing graphs
  • Joining the dots on a line graph when there is no
    evidence of continuity.
  • On a graph of sales for each type of good this is
    obvious.
  • On a graph of sales vs. time its not so obvious.
  • Claiming a trend from 2 or 3 data points.

15
Exercises
  • Using these data, decide the best type of graph
    to use. Sketch the graph. What does it tell you?
    How would you label the axes?

16
Useful links
  • http//www.spreadsheetconverter.com/excel-chart-we
    b.htm
  • http//office.microsoft.com/en-us/excel/HA01034607
    1033.aspx
  • http//www.andypope.info/charts.htm
  • (VBA code)
Write a Comment
User Comments (0)
About PowerShow.com