Dataset Tutorial

From Linked Earth Wiki
Revision as of 19:43, 28 June 2016 by Khider (Talk | contribs) (Edit the tutorial page with current use rights, and updating links.)

Jump to: navigation, search

This page describes a tutorial for using the Linked Earth wiki. The tutorial is structured in an incremental way, getting more complex as you progress. There are four main levels of difficulty, which allow you to perform different tasks in the Linked Earth wiki:

  • Trivial level: how to request an account and log into the wiki.
  • Basic annotation level: how to upload LiPD datasets and update basic metadata fields.
  • Advanced annotation level: how to create and reuse properties for annotation, for instance dataset location annotation.
  • Advanced wiki functionality: create and edit wiki pages, special pages, and uploading materials (pdfs, images).

Finally, we list a set of best practices for creating wiki materials.


Trivial level

Getting a Linked Earth wiki account

To get an account in the Linked Earth wiki, fill in the membership form.

Logging in the Wiki

To log in, just click on the link at top right of the wiki and insert your user/password as shown on Figure 1.

Figure1: Log in form

Basic annotation level

Uploading a LiPD file

After logging in, use this link to upload a LiPD file: http://wiki.linked.earth/Special:WTLiPD, a special page for managing datasets already in the LiPD format. Select the browse button and choose the .lpd file you want to upload as shown in Figure 2:

Figure2: Manage LiPD Data page

The page will be created and your dataset will appear at the top of the "Current LiPD Dataset list" on the Main Page. By clicking on the dataset link, you will be able to see the automatically extracted metadata from the LiPD file, as shown in Figure 3. New "crowd" properties will be automatically created from the LiPD file is these properties are not in the current core ontology.

Figure2: Wiki page created automatically from the metadata of the LiPD dataset

Congratulations! Your LiPD file has been successfully added to the Linked Earth wiki.

Exercise: upload your own LiPD file through the "manage dataset" page in the wiki.

Annotation of a LiPD file

Once a LiPD file is uploaded, the metadata about that file is shown in a table with two columns. The column on the left of each table contains the properties describing the LiPD file, while the column on the right states the value associated with each property. Figure 3 shows an example, where the "archive type" of CAN9Neukom2014 is "Tree".

Now that the metadata about the file is published, we have to review whether it is correct or not, and edit it if appropriate. For example, the first thing we notice for the CAN9Neukom2014 dataset in Figure 3 is that there are no investigators associated with the dataset. In order to edit a property value (whether existing or not defined) we have to click on its corresponding row as indicated in Figure 4. For instance, in Figure 4 we can add "Daniel" as an investigator.

Figure 4: Editing a property value for a LiPD file

Each annotated values can be edited or removed. If we want to remove "Daniel" from the investigators of the dataset, click on the row and on the red cross button of the left as shown in Figure 5:

Figure 5: Deleting a property from a dataset

Each property/property value added to the page is tracked on the page history. The page history is accessible through the "View history" button at the top of any wiki page (see Figure 1 for more details). This way, everyone can check the edits done by other wiki users. Figure 6 illustrates the edits done to the CAN9Neukom 2014 dataset: the latest change added a propertyValue adding "Daniel" as an investigator.

Dataset contributors can only edit the properties associated with the dataset they contributed. Starting at the Basic Editor level, users can edit datasets contributed by other users. If you would like to become a basic editor, please [earth@gmail.com email the Editorial Board.

Figure 6: History of the edits done to the CAN9Neukon2014 dataset

If there is a disagreement between two researchers, a discussion may be started on the "Discussion" page, as depicted on Figure 7. Contributors and Editors may edit any Discussion Pages on the wiki.

Figure 7: Creating a discussion page on CAN9Neukon2014 dataset

Each contribution deserves credit. The wiki pages of the Linked Earth wiki track the authors who have contributed automatically. Figure 8 shows the "Credit" table, at the bottom of the wiki page.

Figure 8: credits of the CAN9Neukon2014 dataset

Dataset versioning

In LiPD all the uploaded datasets should follow a x.y.z notation, where "x" refers to important changes in the dataset's metadata (e.g., the creation of a new age model using a different code), "y" refers to changes to the data following a publication (e.g., adding data further back in time without changing the model underlying the interpretation) and z refers to minor changes not associated with a publication (e.g., typos). For example, the first official release of a dataset would be 1.0.0. If I fix a small typo, I would create version 1.0.1.


Exercise: Annotate the version of the dataset in the recently created page, following the x.y.z notation and using the property "datasetVersion".

Concept annotation in a LiPD file

As shown in Figure 4 and Figure 5, some of the annotated values like "Tree" already have links to other pages. These pages can be further populated and edited by domain experts. To do so click on the Edit tab at the top of the page as shown on Figure 9 for the "Tree" archive. The article was created as a stub, awaiting domain experts' contribution, further gathering field knowledge.

To learn more about Wiki editing, visit the MediaWiki Help.

By linking to other existing pages we can connect different LiPD datasets (e.g., if two different datasets are of the same archive type, they will link to the same page).

When a link appears in red, it means that the page has not been defined yet. Anyone can click on these red links to associate a new page with its documentation. An example can be seen on Figure 10.

Figure 9: Tree concept definition on the Linked Earth wiki
Figure 10: Creating a page for an unexisting concept (highlighted in red)

Exercise: edit the page "Tree ring width", which does not have a definition at the moment, and add a test definition. Use this link as the "Archive Type" value on your uploaded LiPD dataset .


Advanced annotation level

Property annotation in a LiPD file

Until now we have covered how to annotate property values and associate them to concepts and existing pages in the wiki. In this next step we will see how to create new properties to describe a dataset, i.e., adding new annotations to our dataset outside of the standard properties shown in Figure 3.

The "Properties" box, placed under the "Standard Properties" table, allows users to edit and create new properties and values. An example is shown in Figure 11. By clicking on the "plus" sign in the title, a new row will appear on the table. The row has two fields, one for adding the property name we want to use to describe the dataset (e.g., title, description, name, etc.) and another row for inserting the property value.

Figure 11: adding a property for the LiPD file in Figure 3

Before adding a new property name, it is important to note that a similar property may already exist in the ontology to describe the same metadata. The properties are case-sensitive. For instance, imagine that we want to add a "description" to the dataset. If we start typing the property, we see that it already exists, and we can select it for our purposes. Selecting existing properties is important, as it helps to structure and control the content uploaded to the wiki.

Exercise: create a "title" property and use it to annotate your uploaded dataset. Check if the property already exists. If it doesn't, create it.

Adding location to a dataset

To automatically add your dataset to an existing query map such as the one on our Main Page, the coordinates need to be provided. To do so, follow the three steps illustrated on Figure 24. The first step consists of linking the location used to collect the data. Any location is valid, from a single point (e.g., the dataset was collected from point x,y of mountain z) to a polyline (e.g., a river), or even a polygon (e.g., a mountain, a city or even a country). In the example, we are linking the dataset to the location "Central Andes composite 9", where the data was collected.

Once the location page has been created, we can annotate its name and its associated geometry with the property "hasGeometry". This property takes into account the fact that a location may change over time (e.g., a river could change its course), and hence the geometry would change without affecting the location itself.

Finally, we add the coordinates to the Geometry (if the page doesn't exist already). For this we use the AsWKT property, which indicates to the system that the coordinates are in Well Known Text format (https://en.wikipedia.org/wiki/Well-known_text). Since in this case we are representing a point, we also add type "Point" as property for the geometry.


Figure 24: Steps for adding location to your dataset.

Adding and extending concepts (e.g., how do I refer to the instrument used in my lab?)

Sometimes one may want to extend some of the concepts that already exist in the wiki. For example, imagine that I have measured a variable of a table (d18Og.rub-w) with the stable isotope ratio mass spectrometer instrument housed in my lab. If I state that the variable d18Og.rub-w was measured by a "stable isotope mass spectrometer instrument" (under the "instrument" category) I would be losing information: are all stable isotope mass spectrometer instruments the same?

We need to state which stable isotope mass spectrometer was used. In this case, the one in my lab.

Hence, we need to create the concept "StableIsotopeRatioMassSpectrometerInMyLab", referring to a specific "Instrument" with its own property values. For instance, two instruments from the same brand/model could have different reported uncertainty. Following the example shown in the previous section, we would need to create the "StableIsotopeRatioMassSpectrometerIfMyLab" page and annotate it with the category "StableIsotopeRatioMassSpectrometer". This category would become a new category in the wiki, referring to stable isotope ratio mass spectrometer in the more general sense.

Exercise: Use an instrument of your lab to describe a variable from a dataset. If the category of your instrument does not exist, create a new instrument category (e.g., stable isotope mass spectrometer).

Advanced wiki functionality

Creating a personal page

The Linked Earth wiki allows to create different types of pages. Personal pages are a special type of wiki pages, and we will use them to illustrate how to create a new wiki page in general. After logging in on the wiki, click on your name. You can find it on the top of any wiki page, as shown in Figure 12:

Figure 12: location of the personal page for a user

Clicking your name will take you to a new page, which you can complete with your personal information, by clicking on any of the "edit options" in the page, as depicted in Figure 13:


Figure 13: editing personal information on your wiki page

And adding your personal information on the form (Figure 14):


Figure 14: form to fill with the user information

After you are done, just hit the "Update" button at the bottom of the page (Figure 15):


Figure 15: editing your personal page on the form.

Exercise: Create your personal page and add a personal website and an "about me" description.


Creating a new wiki page

Just go to http://wiki.linked.earth/New_Page

Replace "New Page" above to the name that you want for the page.

Then, either select a category for this page (Figure 16):

Figure 16: Selecting a category for a new page

or, just click on the "Create" link to create a page without any category (Figure 17)

Figure 17: create page button, located on the bottom right of the page

Alternatively, you can search for the page you are looking for (remember that the wiki is case sensitive). If the page does not exist, you will be prompted to create it.

Deleting an existing wiki page

Go to http://wiki.linked.Earth/Name_of_Page Replace "Name of Page" above to the name of the page to delete, as shown in Figure 18.

Figure 18: Wiki page option menu

Then click on the "Delete" link, and delete the page. Figure 19 shows an example:

Figure 19: Deleting menu

Searching existing wiki pages

Before creating any page, it is recommended to search if they exist. Searching a page can be done by entering the terms on the search bar located on the top right of any page (see Figure 20).

Figure 20: Search bar for finding existing wiki pages

Any page containing the word introduced in the search bar will be returned.

Renaming wiki pages

Wiki pages may need to be renamed. For example, due to typos in the page name or community agreement. In order to rename an existing page without losing any of its contents, you should click on the "Move" button under the "More" menu on the top of the page. An example is shown in Figure 21.

Figure 21: Moving an existing wiki page

By clicking on the Move button, the page showed in Figure Figure 22 will ask for the new page name, as well as a reason of the change.

Figure 22: Moving a wiki page form.

After hitting the "Move page" button, your page will be renamed.

Uploading images and files

If you want to add an image or document to your a wiki page (Not a LiPD file, use the special page described in this tutorial to upload your LiPD file), click on the "upload files" button that can be seen on the left of any wiki page (see Figure 23).

Figure 23: uploading an image or document to the Linked Earth wiki.

After you are done selecting the document you want to upload, just hit the "Upload file" button at the bottom of the page. Now you may reference to this file using brackets (e.g., [ [ File:File.png ] ]). You can see the different options for showing files in this page: https://www.mediawiki.org/wiki/Help:Images


Best practices

This section will cover naming convntions for instances, publications or versions in a dataset. This is currently under discussion.

TBD