Difference between revisions of "Gnuplot/C2/Error-bars-and-data-fitting/English"

From Script | Spoken-Tutorial
Jump to: navigation, search
Line 33: Line 33:
 
'''Pre-requisites'''
 
'''Pre-requisites'''
 
[https://spoken-tutorials.org https://spoken-tutorials.org]
 
[https://spoken-tutorials.org https://spoken-tutorials.org]
|| To follow this tutorial, * Learner must be familiar with the basics of '''gnuplot'''.  
+
|| To follow this tutorial,  
 +
* Learner must be familiar with the basics of '''gnuplot'''.  
 
* For the prerequisite tutorials, please visit this site.
 
* For the prerequisite tutorials, please visit this site.
  
Line 39: Line 40:
 
|| '''Slide Number 5'''
 
|| '''Slide Number 5'''
 
'''Code files'''
 
'''Code files'''
|| The input file used in this tutorial is provided in the '''code file''' link.
+
||  
Please download and extract the files.
+
* The input file used in this tutorial is provided in the '''code file''' link.
 +
* Please download and extract the files.
  
 
|-
 
|-
Line 50: Line 52:
 
|| Hover mouse over '''x''', '''y''' and '''dy''' columns on the screen for the file.
 
|| Hover mouse over '''x''', '''y''' and '''dy''' columns on the screen for the file.
 
|| The first column is x data.
 
|| The first column is x data.
 +
 
The second column is '''y''' data.  
 
The second column is '''y''' data.  
 +
 
The third column is error in the '''y''' data.
 
The third column is error in the '''y''' data.
  
Line 62: Line 66:
 
|| Enter the command '''cd Desktop '''.
 
|| Enter the command '''cd Desktop '''.
 
Enter the command '''gnuplot'''.
 
Enter the command '''gnuplot'''.
 +
 
|| Change the directory to '''Desktop''' and open '''gnuplot'''.
 
|| Change the directory to '''Desktop''' and open '''gnuplot'''.
  
Line 83: Line 88:
 
|| Hover mouse over '''yerrorbars'''.
 
|| Hover mouse over '''yerrorbars'''.
 
|| If error limits are on the '''x''' data, we have to use, '''x''' errorbars term for plotting.
 
|| If error limits are on the '''x''' data, we have to use, '''x''' errorbars term for plotting.
 +
 
The graphic window opens.
 
The graphic window opens.
  
Line 96: Line 102:
 
|| Hover mouse on the graph.
 
|| Hover mouse on the graph.
 
|| The points may deviate from an accurate exponential decay.  
 
|| The points may deviate from an accurate exponential decay.  
 +
 
This could be due to measurement errors.
 
This could be due to measurement errors.
  
Line 113: Line 120:
 
* Make initial guess values for the coefficients in the equation.
 
* Make initial guess values for the coefficients in the equation.
 
* Optimal values for the coefficients is found by an iterative process.
 
* Optimal values for the coefficients is found by an iterative process.
 
  
 
|-
 
|-
Line 121: Line 127:
 
||  
 
||  
 
* Check for goodness of the fit
 
* Check for goodness of the fit
* A good fitting is measured by a low value of chi square  
+
* A good fitting is measured by a low value of chi square and
 
+
* Display the fitted data with the starting dataset
 
+
 
+
and * Display the fitted data with the starting dataset
+
  
 
|-
 
|-
Line 131: Line 134:
 
and press '''Enter'''.
 
and press '''Enter'''.
 
|| First, let's define the function.
 
|| First, let's define the function.
 +
 
In the '''gnuplot''' prompt, type f of x is equal to a times e to the power minus k x.
 
In the '''gnuplot''' prompt, type f of x is equal to a times e to the power minus k x.
  
Line 155: Line 159:
 
'''k=0.5'''
 
'''k=0.5'''
 
|| Enter commands to set initial guess values.  
 
|| Enter commands to set initial guess values.  
 +
 
Set '''a''' to 150 thousand and '''k''' to 0.5 .
 
Set '''a''' to 150 thousand and '''k''' to 0.5 .
  
Line 161: Line 166:
 
and press '''Enter'''.
 
and press '''Enter'''.
 
|| To fit the data type the command,
 
|| To fit the data type the command,
 +
 
'''fit f of x''' in single quotes the file name.
 
'''fit f of x''' in single quotes the file name.
 +
 
Here it is '''error dash bar dot txt'''.  
 
Here it is '''error dash bar dot txt'''.  
 +
 
Then '''using 1 colon 2 colon 3''' via a comma k .
 
Then '''using 1 colon 2 colon 3''' via a comma k .
 +
 
Here '''a''' and '''k''' are the '''coefficients'''.
 
Here '''a''' and '''k''' are the '''coefficients'''.
  
Line 176: Line 185:
 
Data fitting is seen on the screen.
 
Data fitting is seen on the screen.
 
|| Press '''Enter''' to run the data fitting algorithm.
 
|| Press '''Enter''' to run the data fitting algorithm.
The coefficients in the equation are optimized by an iterative process.  
+
 
 +
The coefficients in the equation are optimized by an iterative process.
 +
 
The output is generated on the screen.
 
The output is generated on the screen.
  
Line 235: Line 246:
 
|| Enter the command, '''plot f(x) lw 2 title 'fitted data', "xydy.txt" using 1:2:3 with yerrorbars pt 7 ps 1.5 notitle''' .
 
|| Enter the command, '''plot f(x) lw 2 title 'fitted data', "xydy.txt" using 1:2:3 with yerrorbars pt 7 ps 1.5 notitle''' .
 
|| Enter the command as seen on the screen, to plot both data together.
 
|| Enter the command as seen on the screen, to plot both data together.
 +
 
We are plotting the fitted equation fit along with the initial data points.
 
We are plotting the fitted equation fit along with the initial data points.
  
Line 241: Line 253:
 
|| The fitted data is represented by a line with a line.
 
|| The fitted data is represented by a line with a line.
 
I am specifying the legend for '''f of x''' as '''fitted data'''.
 
I am specifying the legend for '''f of x''' as '''fitted data'''.
 +
 
No legend title is added for the starting data set as no '''notitle''' is  
 
No legend title is added for the starting data set as no '''notitle''' is  
 
mentioned.
 
mentioned.
Line 248: Line 261:
 
Show symbol with error bar.
 
Show symbol with error bar.
 
|| Data points are represented by symbols with '''error bar''' and without a line style.
 
|| Data points are represented by symbols with '''error bar''' and without a line style.
 +
 
I have specified a filled circle '''symbol''' and 1.5 for point size.
 
I have specified a filled circle '''symbol''' and 1.5 for point size.
  

Revision as of 13:44, 14 January 2020

Visual Cue Narration
Slide Number 1

Title Slide

Welcome to the tutorial on

Error bars and data fitting in gnuplot.

Slide Number 2

Learning Objectives

In this tutorial, we will,
  • Learn to add error bars in a plot
  • Learn about data fitting
  • Write an equation to fit the data
  • Make initial guess for value of the coefficients
  • Fit the dataset to the equation
  • Draw an arrow object in the graph
Slide Number 3

System and Software Requirement

To record this tutorial, I am using,
  • Ubuntu Linux 16.04
  • Gnuplot 5.2.6 and
  • Gedit 3.18
Slide Number 4

Pre-requisites https://spoken-tutorials.org

To follow this tutorial,
  • Learner must be familiar with the basics of gnuplot.
  • For the prerequisite tutorials, please visit this site.
Slide Number 5

Code files

  • The input file used in this tutorial is provided in the code file link.
  • Please download and extract the files.
Go to Desktop. Go to Desktop.

I have a, x, y, y-error type data, in a text file.

Hover mouse over x, y and dy columns on the screen for the file. The first column is x data.

The second column is y data.

The third column is error in the y data.

Close gedit.

Press Ctrl+Alt+T.

Close the file.

Open a terminal.

Enter the command cd Desktop .

Enter the command gnuplot.

Change the directory to Desktop and open gnuplot.
Press Ctrl+L. I will clear the screen.
Cursor on the terminal. Let's plot the data to see the trend.
Enter the command plot 'error-bar.txt' using 1:2:3 with yerrorbars . Enter the plot command as seen on the screen.
Hover mouse over :3 . The command, colon 3 with y error bars adds the error to the plot.
Hover mouse over yerrorbars. If error limits are on the x data, we have to use, x errorbars term for plotting.

The graphic window opens.

Hover mouse over decay curve. Let's fit this graph to an equation.
Cursor on the graph. Here, the data points are likely to follow an exponential decay.
Hover mouse on the graph. The points may deviate from an accurate exponential decay.

This could be due to measurement errors.

Point mouse next to data. We will fit the given data points to an exponential decay function.
Cursor on the screen. Let's see a few steps involved in fitting data points with an equation.
Slide Number 6

Steps for Fitting Data

  • Define an equation to represent the data.
  • Make initial guess values for the coefficients in the equation.
  • Optimal values for the coefficients is found by an iterative process.
Slide Number 7

Steps for Fitting Data

  • Check for goodness of the fit
  • A good fitting is measured by a low value of chi square and
  • Display the fitted data with the starting dataset
Type, f(x) = a * exp(-k*x)

and press Enter.

First, let's define the function.

In the gnuplot prompt, type f of x is equal to a times e to the power minus k x.

Go to graphical window. Make an educated initial guess for the values of a and k.

Go to the graphical window.

Hover mouse near top of y axis. From the graph, I will place the initial value of a at 1,50,000.
Hover mouse to show ½ decay. For an exponential decay, I will place the initial guess of k around 0.5 .
Close the graphical window. Close the graphical window and go to the gnuplot terminal prompt.
Enter the commands,

a=150000 k=0.5

Enter commands to set initial guess values.

Set a to 150 thousand and k to 0.5 .

Type, fit f(x) ‘xydy.txt’ using 1:2:3 via a,k

and press Enter.

To fit the data type the command,

fit f of x in single quotes the file name.

Here it is error dash bar dot txt.

Then using 1 colon 2 colon 3 via a comma k .

Here a and k are the coefficients.

Hover mouse over :3 . We could leave out the colon 3 in the command.

Then, errors in the y data are not considered during the data fitting process.

Press Enter.

Data fitting is seen on the screen.

Press Enter to run the data fitting algorithm.

The coefficients in the equation are optimized by an iterative process.

The output is generated on the screen.

Scroll up. Let's scroll up the screen.
I see an error warning message on the top.
Hover mouse over the chi square table. Notice a table with chi square and new values of a and k after each iteration.
Incorporate the number of iterations. The program reports that, the fitting process converged after 10 iterations.
Hover mouse over final sum of square of residuals. Many fitting parameters are reported in the output.

Program reports, final sum of square of residuals.

Hover mouse over iteration value. Notice, relative change in values after the last iteration.


The number is very small.

Hover mouse over Degrees of freedom. Degrees of freedom is 9.
Hover mouse over RMS. RMS of residuals is around 0.17.
Show a and k values on the screen. Updated values of a and k and their error in estimation is also shown.
Hover mouse over the correlation matrix. The correlation matrix of variables is at the end of the output.
Cursor on the terminal. Now we have a function that fits the given data points.

Let's plot the data points and function together.

Press Ctrl+L . I will clear the screen.
Enter the command, plot f(x) lw 2 title 'fitted data', "xydy.txt" using 1:2:3 with yerrorbars pt 7 ps 1.5 notitle . Enter the command as seen on the screen, to plot both data together.

We are plotting the fitted equation fit along with the initial data points.

Hover mouse near the legend. The fitted data is represented by a line with a line.

I am specifying the legend for f of x as fitted data.

No legend title is added for the starting data set as no notitle is mentioned.

Show the line on the graphics window.

Show symbol with error bar.

Data points are represented by symbols with error bar and without a line style.

I have specified a filled circle symbol and 1.5 for point size.

Enter the command, set xrange [0.95:5.05] . Let's also set x axis limits with set xrange command as seen.
Type replot and press Enter to see the graphics window. Replot to see the results.
Show this web site address screen shot.

http://gnuplot.sourceforge.net/demo/fit.html

Check the gnuplot website for example scripts on data fitting.
Cursor on the screen, around the (1.3,6000) region. Let’s draw an arrow object in the graph.

I want to draw an arrow next to a point in the graph to highlight it.

Move the mouse on the screen to show the arrow placement around (1.63,77200 to 1.7, 62000). In the graph, I want to add an arrow between these two points.
Highlight in video, the co-ordinates seen on the graph window.

(1.63,77200 to 1.7, 62000).

I will note down the coordinates as seen in the graphics window.
Go to the terminal. Go to the terminal.
Type, set arrow 1 from 1.63,77200 to 1.7, 62000 and press Enter. Enter the commands as seen on the screen.

Here, one is the name I have given for the arrow object.

Type replot and press Enter. Replot to see the updated result.
Cursor on the graphics window. I will also add a second arrow in another direction.
Type,

set arrow 2 from 1.63,77200 to 1.4, 62000 and press Enter.

Go to the terminal and enter the command as seen on the screen.

I will name this arrow as second.

Type replot and press Enter. Replot to see the updated result.
Cursor on the graphics window. Now we see two arrows in the graphics window.

To remove the object, we have to unset the object.

Type unset arrow first and press Enter. Go to the terminal.

Enter the command, unset arrow first to remove the first arrow.

Type replot and press Enter. Replot to see the updated result.
Cursor on the graphics window. The arrow that was named first is now removed from the graphics window,
Close the graphics window.

Type quit and press Enter.

Close the graphics window and quit gnuplot.
Slide Number 8

Summary

To summarize, in this tutorial, we learned to
  • Incorporated error bars in a plot graph
  • Fit a given set of data points to an equation
  • Plotted the fitted curve along with parent data
  • Added and removed an arrow object
Slide Number 9

Assignment 1

For the assignment activity, please do the following.
  • For the data file assignment.txt, make an xy graph with xyerrorbars in the graph.
  • This file is available in the Code files link.
Slide Number 9

Assignment 2

  • Fit the data to a double exponential decay curve with yerrorbars.
  • Plot the fitted data
  • Draw an arrow object in the graph at the position of your choice.
Slide Number 10

Spoken Tutorial Project

This video summarises the Spoken Tutorial Project .

Please download and watch it.

Slide Number 11

Spoken Tutorial workshops

We conduct workshops and give certificates.

For more details, please write to us.

Slide Number 12

Forum for specific questions:

Post your timed queries in the forum.
Slide Number 13

Acknowledgement

Spoken Tutorial Project is funded by MHRD, Government of India.
This is Rani, from IIT, Bombay.

Thank you for joining.

Contributors and Content Editors

Madhurig, Ranipv076, Snehalathak