DataPilot example 3: Frequency distribution
For showing the frequency of incidents, Calc uses the function FREQUENCY. This formula has to be used in a so-called matrix formula. This is an advanced feature. Alternatively you can use the DataPilot, which requires almost no knowledge of spreadsheets.
In our example we want to investigate the number of emails that go to the Germanophone support mailing list (email@example.com). We want to know how the activity on the list is distributed during the day.
The starting point for this example is the mailbox file of the Thunderbird mail client. The command
grep 'Delivery-Date:' Inbox
creates a text file that contains for the timespan of more than 2 years one line for each message, with date and time.
To import the data into a table in Calc, open the text file users.txt with the raw data. The Text Import dialog appears; here you can choose your import options. Use the following properties:
Separator options: Fixed width
Divide into columns at position 20, 31 and 40
Column 1: Hide
Column 2: Date (DMY)
Column 3: Standard
Column 4: Hide
The figure below shows the imported raw data with a line added at the top containing titles for each column.
Solution with a matrix formula
To calculate the frequency you have to create 24 classes, one for each hour. In the next column you enter the number of emails with the function FREQUENCY.
The first argument is the cell area with the times of all approximately 38,000 emails. The second argument is the cell area E2:E25 that describes the frequency classes. To enter the formula, first select the cell area F2:F25, then enter the formula. Then finish the formula by using the key combination Shift+Ctrl+Enter. This indicates to the program that you want to use a matrix formula. To indicate the matrix formula, the program uses curly brackets.
This technique is most often known only to advanced users.
Solution with the DataPilot
With the DataPilot you can achieve the same result much easier and faster. The solution is also possible for less advanced users. Starting with the raw data, you need only a few mouse clicks:
- Select the cell A1 (or any other cell within the list).
- Choose Data > DataPilot > Start and click OK.
- In the DataPilot:
- Drag time into the Row fields area.
- Drag date into the Data fields area.
|This may be a very time-consuming process because of the large number of items. The time does not depend that much on the number of lines but rather on the number of rows needed for the table that contains the results.|
The next figure shows the Data Field dialog for the data field Number – Date. Click More and select as type % of column.
The result is shown below with absolute occurrence on the left.
Whether the relative values are shown as a decimal (0.1) or as a percentage (10%) depends only on the cell formatting itself and has nothing to do with the DataPilot.
|Content on this page is licensed under the Creative Common Attribution-Share Alike 3.0 license (CC-BY-SA).|