Skip to content

Latest commit

 

History

History
259 lines (155 loc) · 15.8 KB

LabExercise3.md

File metadata and controls

259 lines (155 loc) · 15.8 KB
Raw

Lab Exercise 3

Instructions

Complete the following lab exercise and submit your answers as a GitHub flavored markdown file to your personal GitHub repository by Febrary 15, 2016.

Finding the Paleobiology Database Website

The URL for the Paleobiology Database is www.paleobiodb.org. However, because you are all honorary members of the development team, you can also use the special development server at www.training.paleobiodb.org. Go there now in your web browser. The first thing that you should see is the SPLASH page.

The Paleobiology Database (PBDB for short) splash page has a lot of information packed into it. At the bottom of the screen you will see some basic stats on the types and quantity of data located in the database.

Data Type Definition
References Scientific articles, books, monographs, or other sources of data.
Taxa A taxon (plural taxa) is a group of one or more populations of an organism or organisms seen by taxonomists to form a unit.
Opinions Different opinions on the correct taxonomic name/identification of different fossil taxa.
Collections A group (collection) of fossil taxa at a specific location.
Occurrences An individual observation of a taxon at a specific location.
Scientists The number of scientists that are officially involved in the Paleobiology Database initiative.

References

All data in the PBDB can ultimatley be traced back to one or more references. The interface for searching and viewing references is currently being overhauled this semester, which is why there is no search button on the splash page. You can still access the old references search feature by clicking here.

The references search page should look something like this.

Exercise Questions 1

Let's take a look at a great scientific paper by Steven M. Holland and Mark E. Patzkowsky.

Use the reference search tool to look up collections associated with this paper and answer the following questions.

  1. How many collections are associated with this references?

  2. What is the reference id number for the article?

Once you have answered the above questions, click the view collections hyperlink to see a print out of the collections associated with the study.

Click on collection no. 72438. Answer the following questions about this collection.

  1. The first taxon in the taxonomic list is Rafinesquina alternata. Next to the taxonomic name is the citation (Conrad 1830), what is the significance of this citation?

  2. What is the class, order, family, genus, and species name of the second taxon in the taxonomic list?

  3. In what County was the data collected?

  4. What age (Period) is the data from?

  5. What is the geologic formation where the data was found?

Collections

Collections are useful for getting additional information about the age, location, and geologic context of collected fossils. They are, however, generally a poor tool for data analysis. This is because there is no standard operational definition of a collection in the Paleobiology Database.

For example, collection 72438 from the above example represents a single sample from the study by Holland and Patzkowsky. In that study, a sample represents a single bedding plane (i.e., the top of a single rock layer) between 100 cm2 and 1600 cm2 in size.

In contrast, collection 91240 represents a single sample in a study by Ivany et al. 2009, Ref# 30540. In that study, a sample was defined as an entire rock outcrop (multiple beds), generally several square meters in extent.

If you blindly compared these two collections, you would be making an apples and oranges comparison.

Occurrences

Occurrences are the number of collections that contain a taxon. Since the size and definition of collections is variable, the meaning of occurrences is also somewhat imprecise.

Therefore, as we progress in this class, you will see that often times the first step of any data analysis project using the PBDB is to reorganize occurrences into a more sensible and standardized format. We will discuss occurrences more when discussing how to download data.

Exercise Questions 2

Return to the SPLASH page, and enter the PBDB navigator tool. This tool is the best way to visualize the age and location of collections in the PBDB.

Look at the search bar prompt in the top right corner. Navigator will allow you to enter a geologic time period, a taxon, an authorizer, or a geologic unit. Let's look for the genus Abra.

  1. Zoom in so that you can see from Texas to Florida and from Florida to New York. You can zoom using the mouse wheel, by double-clicking, or clicking the + and - signs. Some of the occurrences are orange and others are yellow, what is the significance of the different colors?

  2. Zoom back out. Add an additional filter into the searchbar, the Ypresian stage. The Ypresian is a time interval ranging from 47.8–56.0 million years ago. In what countries are there Ypresian occurrences of Abra?

  3. Clear the Abra and Ypresian filters from the search. Look for the genus Ambonychia. Within the United States find the city with the most occurrences of Ambonychia. What is the name of this city?

  4. What age (Period) are most Ambonychia occurrences?

Add in your answer to question 4 as an additional filter. Click on the little icon of South America breaking away from Africa on the left side of the screen. This icon rotates the continents back to their position in the specified time-period. Note that it requires you to have set a specific time-period as a filter.

  1. During this time-period, were most occurrences of Ambonychia arrayed parallel or perpendicular to the equator?

  2. Click on the little insect icon on the left side of the screen. This brings up taxonomic information on the target taxon. What order does Ambonychia belong to?

Downloading Data

You can download the data displayed in your Navigator window using the little arrow icon on the left side of the screen, but its options are limited.

To customize the data you want, use the new and more detailed download form. To find the form, return to the SPLASH page and click on Download Data. This download form uses the new Paleobiology Database API. Once you are more advanced, you will be able to download data directly into R using the API, and will no longer need to use Navigator or the download form.

Let's try downloading all collections of both Ambonychia and Abra as a tab-separated file.

  1. Select Collections
  2. Select Tab-separated values (tsv)
  3. Enter Abra, Ambonychia into the Taxon textbar.

If you were successful you should have gotten a blue URL, describing your data request.

https://paleobiodb.org/data1.2/colls/list.tsv?datainfo&rowcount&base_name=Abra,Ambonychia

Exercise Questions 3

For the following questions generate the appropriate URL for the following data queries.

  1. What is the appropriate URL for downloading all occurrences of Ambonychia in the Lexington Limestone as a JSON?

  2. What is the appropriate URL for downloading all occurrences of mammals present in the Paleocene through Oligocene epochs as a csv?

  3. What is the appropriate URL for downloading all opinions on the order Testudines in the Mesozoic?

  4. What is the appropriate URL for downloading all collections of Aves, Marsupialia, and Sirenia in the United States as a csv?

  5. What is the approopriate URL for downloading all occurrences of the gastropod genus Ficus as a csv (Hint: There is also a plant genus named Ficus)?

Exercise Questions 4

The next set of questions is free form, in that you can find the answer to the following questions using any of the PBDB tools discussed so far.

  1. What family does the genus Gastrocopta belong to?

  2. There is only once occurrence of Isoetes in Portugal. What age is it?

  3. What is the age of the oldest occurrence of Gastrocopta?

  4. There is only one occurrence of Tiktaalik in the Paleobiology Database? Was that occurrence located in the tropics or the extratropics when it was alive?

  5. There are two occurrences of Namacalathus in Sibera. What geologic formations are they found in?

Paleobiology Database API

The acronym API stands for Application Programming Interface. Technical definitions aside, it is a way for users to access data stored in an online database through web addresses (URLs). Companies that store a lot of data (e.g., Google, Twitter, Facebook) make API's available so that 3rd party developers can use their data to make applications. For example, if you've ever played a Facebook game (e.g., Candy Crush, Farmville), those programs were accessing information about you and your friends through the API.

The best way to think about using an API is to imagine it as a map to all the data stored online. You need to use this map to give the computer directions on how to find the particular data you want and access it. When we give directions to a location in the real world, we generally do so in two ways. We either give geographic coordinates (i.e., latitude, longitude, elevation) that specify the destination, or a set of routes to get somewhere (e.g., Take I-90 E to Chicago, then I-80 W to Joliet).

When we access data in R via subscripting (Object[ ]), we are using a coordinate system to point out the data in our object. In contrast, when we access data through an API we are defining a route. In fact, route is the formal terminology. Depending on the size of an API there may be dozens of routes, which may feel overwhelming at first. However, remember that a car map has thousands or hundreds of thousands of roads, most of which you will never travel upon, but you still know how to use a map. It is the same way with an API.

Let's deconstruct a specific API query (i.e., URL).

https://paleobiodb.org/data1.2/occs/list.csv?base_name=Smilodon&interval=Pleistocene

The following figure deconstructs each element of this query.

The first half of the query, before the ? is fairly straightforward because there are only a few possible variations. However, the parameters that come afterwards can become quite cumbersome because there are many varieties of them, and many of them will change depending on what type of data you are using (i.e., collections vs. occurrences). You will need to use the documentation to see a full list of the possible paramters.

You can also access the API documentation from the main SPLASH page.

This will take you to a page that lists the different data routes. If you click on those routes, it will take you to pages that describe different parameters associated with the chosen route. Let's take a breather though and answer some questions.

Exercise Questions 5

https://paleobiodb.org/data1.2/colls/list.csv?base_name=Mammut&interval=Pliocene

  1. In Lab Exercise 2 you downloaded a csv file of ammonite sizes from a github URL directly into R. What code would you use to download the above PBDB data directly into R?

  2. Download the above data into R. What are its dimensions?

  3. Did the above call use the occurrences, collections, references, opinions, or specimens route?

  4. What genus is being called for? What is its colloquial name? What age did I limit the data query too?

  5. Look through the service documentation for the appropriate route (based on your answer to Question 2). Find out how to extend the age search to range from the Miocene Epoch through to the Pleistocene Epoch. Give the new data query URL.

  6. I want the data query to show me the paleocoordinates (i.e., paleolatitude and paleolongitude) of each data point. Give the updated data query URL.

Writing your own API function in R

Wouldn't it be really convenient if instead of typing out the URL every time, you could write an R function that takes a specific taxon name and interval and downloads the data into R automatically?

Specifically, your final question for this lab is to write a function in R that will take as its arguments a taxon name and an interval, and download all fossil occurrences from the PBDB as a CSV.

Your final product should look like this:

# Download all instances of the genus Abra from the Pleistocene interval
AbraData<-downloadPBDB(taxon="Abra",interval="Pleistocene")

# Your output should look like this
AbraData[1:6,1:6]
  occurrence_no record_type reid_no flags collection_no identified_name
1         94761         occ      NA    NA          7108   Abra aequalis
2        256368         occ      NA    NA         20604   Abra aequalis
3        256386         occ      NA    NA         20606   Abra aequalis
4        425385         occ      NA    NA         41501   Abra aequalis
5        427341         occ      NA    NA         41705   Abra aequalis
6        427901         occ      NA    NA         41740   Abra aequalis

# Download all instances of the genus Tyrannosaurus from the Mesozoic
TRexData<-downloadPBDB(taxon="Tyrannosaurus",interval="Mesozoic")

# Your output shoud look like this
TRexData[1:6,1:6]
  occurrence_no record_type reid_no flags collection_no       identified_name
1        139292         occ   22878    NA         11917     Tyrannosaurus rex
2        139293         occ      NA    NA         11918 Tyrannosaurus cf. rex
3        219998         occ      NA    NA         22654     Tyrannosaurus rex
4        220009         occ      NA    NA         22657     Tyrannosaurus rex
5        280101         occ      NA    NA         26760     Tyrannosaurus rex
6        291021         occ      NA    NA         14640 Tyrannosaurus cf. rex

In order to achieve this you will need to use the paste( ) function. Here are some examples of the paste function in use. See if you can figure out how it fits into the problem.

# Example 1
paste("We","Love","R",sep=" ")
[1] "We Love R"

# Example 2
paste("We","Love","R",sep="")
[1] "WeLoveR"

# Example 3
paste("We","Love","R",sep="!")
[1] "We!Love!R"

# Example 4
LoveR<-c("We Love R")
HateR<-c("We Hate R")
paste(LoveR,HateR,sep=" >> ")
[1] "We Love R >> We Hate R"

Question 6

  1. Write an R function that will take a taxonomic name (as a character string) and an interval (as a character string) as its argument, and will download all fossil occurrences into R. See above.