How to use LibreOffice ?

Here are some tips and elements of method when creating graphs in LibreOffice. The latter must be
included inside of your report.

1 - What is the point of a graph ?

When you make measurements during an experiment, you collect raw data. Unfortunately, it is hard to
find properties by looking at tables of numbers. In order to solve this problem, we distribute the data
points in space in order to visualize the properties. This is what we call a graph.

2 - Graphics main goals

Here is a brief recap of the goal of a graph you create in order to illustrate the results o an experiment.

With the advent of computers, it is no longer conceivable to present graphs drawn quickly, not re-
specting position of points or scales,... Then it is always better to use computers. Several software
are possible, but I recommand to use LibreOffice (free and open source) or Excell (accessible through
Strasbourg inscription).

Several remarks :

• Leave only relevant graphics in the report. It is not up to the professor to sort through the data
you provide.
• I recommend to always put the graph near the analysis and the explanations you give about it.
Otherwise, create an appendix and label your graph so that you can easily refer to it.
• The graphics must be correctly presented : title, name of the axes and units, relevant scale.
• Never put line beetwen your data point on graph. It would mean that you know what the signal
looks like in between and we are never sure about that.

• The markers used to indicate data points should always be "+" crosses. It allows exact and easy
reading of the values you found.

3 - How to trace a graph in LibreOffice ?

1. First you need to collect your data in form of a table. Don’t forget to put names above the columns
or below the lines to ensure that you don’t forget what it id. And then select the data you want to
plot.

2. Create a graph by clicking on the diagram menu button

1
3. Then appears the diagram menu. You will have to select the scatter diagram by clicking on

4. Different options will appear for the graph. Select the first one, with no lines. Indeed, we cannot
assume what does the data looks like between our measurement points. Then click on next button

5. There different options are suggested. The ones that interest us really are these ones.

It asks about the disposition of our data. In our example case, data is in column, we should select
this option. But in other PWs, it is asked to proceed data in rows. In that case you should select
the other option.
6. After clicking on the next button, you will see here a new window.

In data series, you can select the data you want to work on. In data ranges, you can change the

data you previously selected for making the graph. For convenience, you can use the button
for selecting with the cursor.
You can also add different curves on the same graph, but we will see that in another section.
7. After configuration of the data, you can click on the next button. Then you find the most revelant
part of the graph ; the legend. Please fill it in the most appropriate way. If it is not done, your
graph will have no purpose.

2
 Before completion After completion
8. Finally you can create your graph using the finish button. Your graph is created !

9. Unfortunately, the graph is created with datapoints marked with squares. It is impossible to know
exactly the value of this datapoint by looking at the graph. We need to replace the squares by
crosses "+". For that, select the graph and enter the modification by double clicking on it. Grey
border should appear on the graph. Then select all the datapoints by clicking on one of them.
Then right click on this point and select Format data series.

10. On the new window that appears, go to the Line menu and choose the "+" cross from the Select...
list and then Symbols. You can later adjust the size and the color of the cross in this same menu if
your graph is too overcrowded.

11. Here is the final result. If you want to modify the axis, enter in the editing mode of the graph
by bouble clicking on it and then double click on the line of the axis. The following menu should
appear. The scale section is very usefull and you can play with it to better highlight your results.

3
4 - How to fit a dataset with a curve and get its equation ?
Now that your graph is ready, you can get information from it. One way of extracting information is
by fitting your plot with a curve. It means that you will ask the software to create a curve based on an
equation type (linear, polynomial, exponential,...) and adjust the different coefficients of the equation of
the curve so that the curve as close as possible to all the points.

1. Enter the editing mode of the diagram by double clicking on it. Select the datapoints and right
click on them. Then cklick on Insert Trend Line.

2. Once you entered the trend line menu, you can select the curve that you think fits the best your
data. In our example it is polynomial function of third order. Select also the Show equation option.
Here is also the result of these manipulations after clicking on OK button.

4
 We can check here that the coefficients are corresponding to the equation we used to plot our graph.

3. Unfortunately the coefficients are very long. This problem can be solved by right clicking on the
equation and select Format Trend Line Equation.

4. In the format menu we will chose the scientific notation for all the coefficients. Here is the final
result.

5 - How to fit data only on a portion of the dataset ?

Sometimes, it is important to represent a whole dataset while only fitting a small section of it. This
section is dedicated to that. In our example, it is possible to fit the data around 0 with a linear fitting.
We will create 3 different curves and only fit the middle one.

5
1. Enter the editing mode and select all the datapoints. Right click and go to the Data ranges... menu.

2. Once in the menu, use the button in order to select the first portion of the curve you don’t
want to be in your fitting.

3. Then add a new curve by clicking on the Add button. And select the following portion you want to
fit using the same technic as previously. Proceed to add a third curve exactly the same way with
the exta bit of the dataset.

4. In the end you should get this result. You just have to format the datapoints the same way by
following the previous procedure to have indistiguishable datapoints.

6
5. You can now select the middle curve and apply the linear fitting following the same procedure as
indicated in the previous section. Here is the result.