2000-01-03

Criteria for accepting charts

The charts are graphical representation of your data. It represent mostly the relationship between two variables x and y where variable x is independent variable and variable y is dependent. In the Microsoft Excel such charts are called - Scatter plot.

There are 3 things related with chart format I will be looking for and ... you should keep in mind.

  1. Chart title. It explains the process or the relationship between variables.
    1. Bad example - "Concentration and absorbance"
    2. Better example - "Concentration versus absorbance"
    3. Best examples -
      "The relationship between concentration (independent value) and absorbance (dependent value)"
      "The dependence of traffic ticket cost and car speed"
      "Absorbance spectrum"
  2. Axis titles. Each axis have its own variable. 
    1. Independent variable on the x-axis
      1. whatever which is predetermined
        1. Time
        2. predefined concentrations
        3. predefined temperature
        4. etc
    2. Dependent variable on the y-axis
      1. Measured quantities
        1. Refractive index
        2. Angles
        3. Voltage
        4. etc.
    3. Axis titles should have physical meaning and unit
      1. Bad examples
        1. T
        2. Sec
        3. "
        4. T,s
      2. Good examples
        1. Time, s
        2. Time [min]
        3. Rotation angle, degrees
        4. Concentration, %
    4. Dimensionless physical quantities can be left without units or their notations can be used
      1. Good example
        1. Refractive index, n
  3. Legend. Chart can have more than one data-set. How to distinguish data between each other?
    1. Bad examples
      1. Series 1, Series 2 ..
      2. Series 1, Linear (Series 1)...
    2. Good examples
      1. Big cylinder, Small cylinder...
      2. Optically active substance, Linear (Optically active substance)
      3. Photospectrometer Sony-XA1
Regression analysis can be applied to certain data-sets. Even more - for one data-set more than one regression types can be applied. Then you can compare which regression type fit the best using R2 value. Use following rules 


  1. Make regression lines distinguishable (choose different colors)
  2. Make regression equation and R-squared value distinguishable (use chosen line color for corresponding equation and R-squared value)
Is everything`s fine in this chart?

Nav komentāru:

Komentāra publicēšana