Data extraction


The video above describes how to perform the data extraction from the articles in your review.


Example

PROJECT chocoholics "Chocoholics in the North America"
CLASSIFICATION
Simple year "Year published" * [1] : int(4) // this is an example
List country "Country" [1] = ["Canada", "Mexico", "U.S.A."]
DynamicList cocoa "Cocoa level" [-1] = ["Bitter", "Bittersweet", "SemiSweet", "MilkChocolate"]

In this example, the data extraction form consists of three categories. In the first one, we collect the publication year of each article we are analyzing. The first keyword is the type of category. Here, Simple means it is a free-form field. The name of the category is year and must be unique. We want to make it mandatory, so we add a *. This means we must collect the year of publication of every article. [1] indicates that only one value is allowed, since there is only one year of publication for each article. Finally, we specify the type of data that is allowed in the category. Here it is an integer number with at most four digits. // are used for comments and are not processed.

The second category collects the country of the study in the article. Since there are only three countries in North America, we can use a List with the three countries. This cannot be modified later on.

The third category collects the cocoa level. [-1] indicates that multiple values are allowed. Since there may be more values added later, we make this category a Dynamic List, with some predefined values. These can be changed later on.

You can then specify a reporting section in the configuration to create charts about the extracted data.


Category types

Here is the list of category types and their options:

Simple

This is for free-form categories where any text can be entered. The format is:

Simple category_name "Label to display" * : type(max) style("pattern") = ["default_value"]
category_name
Variable that uniquely identifies this category.
Label to display
Text to display for this category in the data extraction form.
*
If a * is present, then the category is mandatory. The data extraction form cannot be saved unless a value is entered for this category. The * is optional, if omitted, then the category is optional.
type
Can be any of the following:
  • bool for a check box: can be either true or false (1 or 0).
  • int for a number which can be negative.
  • real for decimal numbers.
  • string is a one-line text.
  • text for multi-line text.
  • date for a date displayed as a calendar.
max
The maximum number of characters the value can have.
pattern
A regular expression to limit the possible values allowed. Click here for more information. The style("pattern") is optional.
default_value
A default value that will appear in the initial data extraction form. It must be between square brackets and quotation marks. It must also be in the correct format depending on type, max and pattern. The ["default_value"] is optional.

For a Simple category, only one value can be entered. If you want to specify multiple values, you can come up with your own notation, such as separating the values with commas. For some categories, like date, this is not possible.

List

This is a category with a predefined list of values. Once the values are defined, it will no longer be possible to change, add, or remove them during the systematic review process. The format is:

List category_name "Label to display" * [n] = ["value 1", "value 2"]
category_name, Label to display, *
Same as the Simple category.
n
The maximum number of values for this category. 0 or 1 means only one value can be selected from the list, 2 means up to two values can be selected, etc. To allow an unlimited number of values, use -1. The number must be between square brackets.
["value 1", "value 2"]
You must enumerate all the possible values of the list here. You can have any number of values in this list. Since this cannot be changed later on, the list of values is mandatory. The list must be between square brackets, each value must be between quotation marks and separated by a comma.

Dynamic List

This is like a List, but the values can be modified during the systematic review process. This is especially useful if you want to share the values in this category between articles in your corpus. It is more flexible than a List and imposes a stronger normalization of the data than a Simple category. This is the most used type of category. The format is:

DynamicList category_name "Label to display" * [n] "Reference name" = ["value 1", "value 2"]
category_name, Label to display, *, n
Same as the List category.
Reference name
Recall that it is possible to change the list values during the systematic review process. To do so, a button called Reference Categories will appear in the menu when you are in the classification mode in ReLiS. Reference name is the name of the button that will be displayed under the menu. You can than add, change or remove the possible values there. This is mandatory. Often, this can be the same as the Label to display.
["value 1", "value 2"]
As opposed to a List, the preliminary list of possible values is optional.

Dependent Dynamic List

This is similar to a Dynamic List, but the values come from another category. This is useful if you want to restrict the possible values, but these values come from another DynamicList category. The format is:

DynamicList category_name "Label to display" * [n] depends_on dependent_category
category_name, Label to display, *, n
Same as the Dynamic List category.
dependent_category
The category_name of another Dynamic List from which the values are populated.

Sub-categories

Any of the above categories can contain sub-categories. A category containing sub-categories is called a super-category. On the data extraction form, adding a value to a super-category will pop-up a secondary form. The format is:

SUPER CATEGORY DEFINITION {
  SUB CATEGORY DEFINITION
  SUB CATEGORY DEFINITION
}

Here, SUPER and SUB CATEGORY DEFINITION are the same as any category type above. All you need is to enclose the sub-categories between curly brackets.

Built-in categories

The following are built-in categories:

note
A text for leaving notes or comments. It is equivalent to
Simple note "Note" [1] : text(500)

You can regenerate the configuration file at any point in time, but beware that some data may be lost if not modified properly.