Row & Document Based

Row-based and Document-based

If you choose row-based or document-labeling as the task type, the goal of labeling is to answer the questions. You can answer the questions in the Document Labeling extension on the right side. (See this Youtube video for visual instructions on how to label row-based projects).

You can navigate to the next question by using your mouse or typing Tab on the keyboard.

You can move to the desired row via the Go menu.

Go to Start will take you to the first row.
Go to End will take you to the last row.
Go to Line will take you to a specific row.
Go to Next Unlabeled Line will take you to the next unlabeled line.
Go to Previous Unlabeled Line will take you to the previous unlabeled line.
Go to Next File will take you to the next file.
Go to Previous File will take you to the previous file.

We not only support displaying all rows for row labeling but also allow you to customize the number of rows displayed on a single page. You can adjust this preference by changing the Number of rows displayed per page in the Preview Settings in Project Creation Wizard Step 2.

Once you set the number of rows preferences in Step 2, you will have a similar interface to the following screenshot and can easily navigate and paginate your labeling process using the Navigation Bar above the table. More information about Table Data here.

We support this Navigation Bar only for Row Based Project

Here are some examples of the Navigation Bar view:

Tabular view

URL view (image)

URL view (website)

You can also navigate to your next document if you upload more than one. On the bottom left side of your labeling interface, there is a navigation bar that you can use to navigate to your next document.

Required question

The asterisk (*) next to the question indicates that the question requires an answer - leaving a required field blank will trigger an error.

Sort and filter column

If you create Text Field, Text Area, Dropdown, Hierarchical Dropdown, Date, Time, Checkbox, Slider, Grouped Attributes and URL question's type, you are able to sort and filter the columns.

For the Text Field, URL and Text Area columns, you can filter by searching the keyword.

For the Dropdown and Hierarchichal Dropdown column, you can filter based on the dropdown value.

For the Date and Time column, you can filter the date range and the time range.

For the Checkbox column, you can filter it based on the true or false value to check whether is it checked or not.

For the Slider column, you can filter it based on the value provided or based on the value range.

For the Grouped Attributes column, you can filter it by clicking the arrow down button besides the question type header name.

It will display a dropdown list of each question type in the Grouped Attributes question type.

You can start to apply a filter based on the labeled answer by clicking the question, and apply the filter.

If you already apply some filters, the arrow down button besides the question header name will be changed into filter icon.

Please note that we haven't supported filter for Radio question inside a Grouped Attributes question type

When the answer type is Dropdown, keyboard shortcuts are displayed in the extension. In the example below, you can click 1 on your keyboard to apply POSITIVE as an answer.

Filter rows

You are allowed to see all rows or the unlabeled rows by clicking the View menu. This feature will help you if your project has many rows.

This feature is only available for in the Reviewer Mode on Row based projects.

Filter Unlabeled row only

Filtering for unlabeled rows allows you to quickly identify and prioritize data points that require labeling. With this option selected, you can easily pinpoint unprocessed data, ensuring no crucial information gets overlooked.

Filter Conflicted only

Handling conflicts in labeled data is essential to achieve accurate results. By isolating conflicted rows, you gain a clear view of data points that have discrepancies in their labels. This empowers you to address conflicts promptly and make informed decisions to resolve them effectively.

Filter Unreviewed only

Efficiently manage your review process by filtering for unreviewed rows. This feature helps you keep track of which data points require verification, allowing you to ensure that all labeled data has undergone the necessary review before being finalized.

Unreviewed only includes both consensus and conflict rows, which essentially means all the rows where the reviewer has not yet submitted an answer.

Filter Reviewed only

If you need to work exclusively with reviewed and approved data, this filtering option is your go-to solution. By focusing solely on reviewed rows, you can confidently utilize accurate and reliable labeled data in your projects.

Hide and rename the headers

You can hide and rename headers by right-clicking on the header.

Mark as complete

Once you have finished labeling, click Mark as complete. This will signify to your team you are done with the project, and it is ready for Review or Export.

Row-based with URL view

There's an option for you to label multiple images by providing the URL of the Images in a column of your Row-based file.

Prepare a Row-based file that contains a URL column

You can store your Images on your preferred storage options (make sure it's accessible). You can also add additional information for each of the images by adding the attributes to the columns (The data can't be edited later).

821B

Datasaur sample - Row (url viewer).csv

Check and set the preview of the row-based labeling

You can set how the media will be previewed on the labeling page. Here are some of the options:

Don't expand: Not previewing image from the URL
Thumbnail: Previewing smaller size of the image from the URL
Large: Previewing the larger size of the image from the URL

Set the viewer setting to URL view

Make sure to change the Viewer Setting from the Tabular View to the URL View. Also, set the URL Columns to the column name of your Row-based file that contains the URLs.

Labeling page of the row-based with URL View

Here you can see your links that are retrieved from the URL that you provided on the Row-based file. Now you can conduct the Row-based labeling with the help of URL View.

See additional information on the document labeling extension

The additional information that are available on the columns of the Row-based file can be found on the Document Labeling extension.

Applying answers for multiple rows

Some users may notice some patterns in their dataset. So, they may need to apply the same answer to multiple rows at once. The good news is Datasaur support that!

Users are allowed to select multiple rows and apply an answer to the selected rows. There are two ways to use this feature:

Select the checkbox that is available per row
Hold ctrl from the keyboard to select multiple rows or you can tick the check box to select multiple rows

After selecting the rows, you can select answers on the questions in Document Labeling extension, then click Submit to apply the answer.

How it works

Select multiple rows that have no answer, then bulk answer
- All rows should have the same answers
Select two rows where one row has an answer and one row has no answer, then bulk answer
- All rows should have the same answers
Select two rows where one row has an answer and one row has no answer, then answer question one and leave the rest
- The same answer should be provided for question one only
If you select multiple rows, there is a possibility that the questions have different answers, hence the Mixed value will be shown below the question:
Case 1: If the user changes the answer value, it will override all answers for the selected question if submitted.
Case 2: Reset button will be shown in case user changes their mind and wish to change back to Mixed value

This capability is only available in the labeler mode.
This capability doesn’t apply for row labeling projects with any of these settings:
- Number of rows per page: 1
- URL viewer

Auto-saved answers

Auto-saved will save your answers as a draft before submitting, and the draft will remain even if you accidentally refresh or close the page. This feature is particularly suitable for row-based or document-based projects that have many questions.

Auto-saved answers apply to both labeler and reviewer. If another reviewer submits an answer in the extension, the current draft will be replaced with the latest answers (compare the timestamps).

Additionally, this feature includes the ability to discard drafts, which can be done by clicking the triple dots in the Document Labeling questions.

Please note that the auto-saved answer feature is only available for single row selection, and this capability also applies to discarding drafts.

For row-based projects, please note that the draft will only be shown in the Document Labeling extension, not in the table
- Answers will be shown in the table once the submission is successful
Changing question sets will remove the auto-saved answers