When two major workshops concluded by the EarthCube CRESCYNT Coral Reef Science and Cyberinfrastructure Network in March 2018, there were some interesting clear outcomes in addition to the practical training and data exploration goals accomplished. The workshops were both structured around Data Science for Coral Reefs. At the end of the first, focused on Data Rescue and data management, participants decided that the most important new topic they learned about was metadata and its uses. At the end of the second, focused on Data Integration and Team Science, people had realized how essential writing good metadata was for being able to make datasets at disparate scales work together well. The metadata lessons were important emergent outcomes, and participants asked that data and metadata experts get together, use the data challenges that arose, and recommend some metadata practices and standards that would work for the coral reef community and its very broad range of data types, repositories, and pre-repository research, storage, sharing and analytical metadata needs.
We were luckily able to do exactly that with one final workshop. Through a jointly staged CRESCYNT-DDH workshop, we pulled together a group of metadata experts, coral reef data managers, representative scientists, and the EarthCube Data Discovery Hub’s scientists and software developers focused on metadata enhancement for finding and using data.
Special guests included Ted Habermann (Metadata 2020 project, co-author of “The influence of community recommendations on metadata completeness”; Stephen Richard (experience with schema.org and metadata standards authoring); coral reef data experts Gastil Gastil-Buhl (Moorea Coral Reef LTER), Hannah Ake (BCO-DMO), and Sarah O’Connor and Zachary Mason (NOAA NCEI’s user metadata writing interface and CoRIS), the three biggest formal repositories for coral reef research data in the US or sponsored by NSF; Eric Lingerfelt, the EarthCube Technical Officer; guests from Scripps; and DDH team members Ilya Zaslavsky, Karen Stocks, Gary Hudman, David Valentine, and Tom Whitenack with their broad and integrative metadata, software, and domain expertise. Ilya and Karen kindly hosted the group at UCSD’s San Diego Supercomputer Center and Scripps Institution of Oceanography.
Important outcomes from the workshop were mutualistic for the two projects. For CRESCYNT, they included cross-mapping an essential set of metadata (as defined by appropriate community repositories) to web standards and producing a draft ISO metadata profile for coral reef data at two levels of dataset access: (1) discovery and sharing (a simpler form with freeform text entry in many of the fields), and (2) understanding and usability at the workbench level (a more detailed form with options to supply more highly specified fields). We will finish writing these and offer them to the coral reef community for feedback and potential adoption.
For the DDH, important outcomes included exploration of the use of the enhanced metadata at different repositories and in science use cases (including the coral reef use case), a deep dive into focusing the future trajectory of the Data Discovery Hub, and some initial planning for an upcoming data science competition that will involve the coral reef data (details to be announced). Read more about DDH and its broader work.
We gratefully acknowledge the generosity of our hosts, travel funding by NSF, the active work and engagement of our participants, and the organizations that allowed their employees time to attend and contribute to this collective effort.
Attendance is free and open to the public, online or in person.
Written while embedded in our CRESCYNT Data Science for Coral Reefs workshops. Amazingly, everyone who participated in workshop 1 – Data Science for Coral Reefs: Data Rescue – learned even more than they thought they would. We’ve had wonderful NCEAS trainers, spectacular participants with amazing datasets, and a lot of hard work over 4 days (March 7-10, 2018).
UPDATE: Here is the Data Rescue workshop agenda we used, with links to all of the training slides.
In the second intensive workshop – Data Science for Coral Reefs: Data Integration and Team Science – people will be introduced to R Studio and GitHub if they have not used them before, and then we will work on exploring techniques for integrating disparate datasets. We’ll start with a pair of datasets at a time, and efforts may involve extracting data from one dataset based on observations from another; upscaling, downscaling, resampling, or summarizing to make intervals and scales mesh – exactly the kind of process that coral reef researchers have said is a recurring challenge in asking bigger science questions.
UPDATE: Here is the Data Integration and Team Science workshop agenda we used, with links to all of those training slides and exercises.
Each workshop group is writing a paper to summarize and share lessons learned, so please stay tuned for those!
We experimented with an unusual process for these workshops: two days of training followed by two days of workathon. We’re liking it! Tell us what you think about these topics and training materials. What other workshop outputs would you like to see?
In preparation for an upcoming Data Science for Coral Reefs: Data Rescue workshop, Dr. James W. Porter of the University of Georgia spoke eloquently about his own efforts to preserve historic coral reef imagery captured in Discovery Bay, Jamaica, from as early as 1976. It’s a story from the trenches with a senior scientist’s perspective, outlining the effort and steps needed to accomplish preservation of critical data, in this case characterizing a healthy reef over 40 years ago.
Enjoy this insightful 26-min audio description, recorded on 2018-01-04.
Transcript from 2018-01-04 (lightly edited):
This is Dr. Jim Porter from the University of Georgia. I’m talking about the preservation of a data set that is at least 42 years old now and started with a photographic record that I began making in Discovery Bay, Jamaica on the north coast of Jamaica in 1976. I always believed that the information that photographs would reveal would be important specifically because I had tried other techniques of line transecting and those were very ephemeral. They were hard to relocate in exactly the same place. And in addition to that they only captured a line’s worth of data. And yet coral reefs are three dimensional and have a great deal of material on them not well captured in the linear transect. So those data were… I was very consistent about photographing from 1976 to 1986.
But eventually funding ran out and I began focusing on physiological studies. But toward the end of my career I realized that I was sitting on a gold mine. So, the first thing that’s important when considering a dataset and whether it should be preserved or not is the individual’s belief in the material. Now it’s not always necessary for the material to be your own for you to believe in it. For instance, I’m working on Tom Goreau, Sr.’s collection which I have here at the University of Georgia. I neither made it nor in any way contributed to its preservation but I’ve realized that it’s extremely important and therefore I’m going to be spending a lot of time on it. But in both cases, the photographic record from Jamaica, as well as the coral collection itself – those two activities have in common my belief in the importance of the material.
The reason that the belief in the material is so important is that the effort required to capture and preserve it is high, and you’ve got to have a belief in the material in order to take the steps to assure the QA/QC of the data you’re preserving, as well as the many hours required to put it into digital format. And believing in the material then should take another step, which is a very self-effacing review of whether you believe the material to be of real significance to others. There’s nothing wrong with memorabilia. We all keep scrapbooks and photographs that we like – things relating to friends and family, and times that made us who we are as scientists and people. However, the kind of data preservation that we’re talking about here goes beyond that – could have 50 or 100 years’ worth of utility.
Those kinds of data really do require them to be of some kind of value, and the value could either be global, regional, or possibly even local. Many local studies can be of importance in a variety of ways: the specialness of the environment, or the possibility that people will come back to that same special environment in the future. The other thing that then is number two on the list – first is belief in the material – second is you’ve got to understand that the context in which you place your data is much more important to assure its survival and utility than the specificity of the data. Numbers for their own sake are numbers. Numbers in the service of science become science. It is the context in which you place your data that will assure its future utility and preservation.
We’re extremely pleased to be able to offer two workshops in March 2018 at NCEAS. The second is CRESCYNT Data Science for Coral Reefs Workshop 2: Data Modeling, Data Integration and Team Science. Apply here.
When: March 12-15, 2018
Where: NCEAS, Santa Barbara, CA
This workshop is recommended for early to mid-career and senior scientists with interest in applying technical skills to collaborative research questions and committed to subsequently sharing what they learn. Participants will learn how to structure and combine heterogeneous data sets relevant to coral reef scientists in a collaborative way. Topics covered on days 1 and 2 of the workshop will cover reproducible workflows using R/RStudio and RMarkdown, collaborative coding with GitHub, strategies for team research, data modeling and data wrangling, and advanced data integration and visualization tools. Participants will also spend 2 days working in small teams to integrate various coral reef datasets to practice the skills learned and develop workflows for data tidying and integration.
The workshop is limited to 20 participants. We encourage you to apply via this form. Workshop costs will be covered with support from NSF EarthCube – CRESCYNT RCN. We anticipate widely sharing workshop outcomes, including workflows and recommendations. Anticipate some significant pre-workshop prep effort.
We’re extremely pleased to be able to offer two workshops in March 2018 at NCEAS. The first is CRESCYNT Data Science for Coral Reefs Workshop 1: Data Rescue. Apply here.
When: March 7-10, 2018
Where: NCEAS, Santa Barbara, California, USA
Recommended for senior scientists with rich “dark” data on coral reefs that needs to be harvested and made accessible in an open repository. Students or staff working with senior scientists are also encouraged to apply. Topics covered on days 1 and 2 of the workshop will cover the basic principles of data archiving and data repositories, including Darwin Core and EML metadata formats, how to write good metadata, how to archive data on the KNB data repository and elsewhere, data preservation workflow and best practices, and how to improve data discoverability and reusability. Additionally, participants will spend approximately 2 days working in pairs to archive their own data using these principles, so applying with a team member from your research group is highly recommended.
The workshop is limited to 20 participants. We encourage you to apply via this form. Workshop costs will be covered with support from NSF EarthCube – CRESCYNT RCN. Participants will publish data during the workshop process, and we anticipate widely sharing workshop outcomes, including workflows and recommendations. Because coral reef science embodies a wide range of data types (spreadsheets, images, videos, field notes, large ‘omics text files, etc.), anticipate some significant pre-workshop prep effort.
Related post: CRESCYNT Toolbox – Estate Planning for Your Data
Data cleaning. Data cleansing. Data preparation. Data wrangling. Data munging.
Garbage In, Garbage Out.
If you’re like most people, your data is self-cleaning, meaning: you clean it yourself! We often hear that 80% of our “data time” is spent in data cleaning to enable 20% in analysis. Wouldn’t it be great to work through data prep faster and keep more of our data time for analysis, exploration, visualization, and next steps?
Here we look over the landscape of tools to consider, then come back to where our feet may be right now to offer specific suggestions for workbook users – lessons learned the hard way over a long time.
The end goal is for our data to be accurate, human-readable, machine-readable, and calculation-ready.
Software for data cleaning:
RapidMiner may be the best free (for academia) non-coding tool available right now. It was built for data mining, which doesn’t have to be your purpose for it to work hard for you. It has a diagram interface that’s very helpful. It almost facilitates a “workflow discovery” process as you incrementally try, tweak, build, and re-use workflow paths that grow during the process of data cleaning. It makes quick work of plotting histograms for each data column to instantly SEE distributions, zeros, outliers, and number of valid entries. It also records and tracks commands (like a baby Jupyter notebook). When pulling in raw datasets, it automatically keeps the originals intact: RapidMiner makes changes only to a copy of the raw data, and then one can export the finished files to use with other software. It’s really helpful in joining data from multiple sources, and pulling subsets for output data files. Rapid Miner Studio: Data Prep.
R is popular in domain sciences and has a number of powerful packages that help with data cleaning. Make use of RStudio as you clean and manipulate data with dplyr and tidyr. New packages are frequently released, such as assertr, janitor, and datamaid. A great thing about R is its active community in supporting learning. Check out this swirl tutorial on Getting and Cleaning Data – or access through DataCamp. The most comprehensive list of courses on R for data cleaning is here via R-bloggers. There’s lovely guidance for data wrangling in R by Hadley Wickham – useful even outside of R.
- Trifacta Wrangler was built for desktop use, and designed for many steps of data wrangling: cleaning and beyond. See intro video, datasheet, demo with Tableau.
- DataCleaner – community or commercial versions; can use SQL databases. Mostly designed for business applications; videos show what it can do.
- OpenRefine gets the legacy spotlight (was Google Refine… now community held). Free, open source, and still in use. Here’s a recent walkthrough. Helps fix messy text and categorical data; less useful for other science research data.
There are some great tools to potentially
steal borrow that started in data journalism:
- Tabula is “a tool for liberating data tables trapped inside PDF files” – extracts text-based pdfs (not scans) to data tables.
- csvkit is “a suite of command-line tools for converting to and working with CSV, the king of tabular file formats.” Helpful for converting Excel to csv cleanly, csv to json, json to csv, working with sql, and more.
- agate is “a Python data analysis library that is optimized for humans instead of machines…. an alternative to numpy and pandas that solves real-world problems with readable code.” Here’s the cookbook.
Finally, Python itself is clearly a very powerful open source tool available for data cleaning. Look into it with this DataCamp course, pandas and other Python libraries, or this kaggle competition walkthrough.
Manual Data Munging. If you’re using Excel, Open Office, or Google Sheets to clean your data (e.g., small complex datasets common to many kinds of research), you may know all the tricks you need. For those newer to data editing, here are some tips.
- To start: save a copy of your original file with a new name (e.g., tack on “initials-mod” plus the current date: YYYYMMDD). Then make your original file read-only to protect it. Pretend it’s in an untouchable vault. Use only your modifiable copy.
- Create a Changes page where you record the edits you make in the order you make them. This also lets you scribble notes for changes you plan to make or items you need to track down but haven’t yet executed (Done and To-Do lists).
- First edit: if you don’t have a unique ID for each row, add a first column with a simple numeric sequence before doing anything else.
- Create a copy of that spreadsheet page, leave the original intact, and make modifications only to the newly copied page. If each new page is created on the left, the older pages are allowed to accumulate to the right (less vulnerable to accidental editing). Name each tab usefully.
- Second edit: if your column headings take up more than one row, consolidate that information to one row. Do not Merge cells. Include units but no special characters or spaces: use only letters, numbers, dashes and underlines.
- Add a Data Definitions page to record your old column headings, your new column headings, and explain what each column heading means. Include units here and also in column headings where possible.
- In cells with text entries, do not use bare commas. Either use semicolons and dashes instead of commas in your text, or enclose text entries in quotation marks (otherwise creates havoc exporting to and importing from csv).
- Add a Comments column, usually at the end of other columns, to record any notes that apply to individual rows or a subset of rows. Hit Save, now and often.
- Now you’re free to sort each column to find data entry typos (e.g., misplaced decimals), inconsistent formats, or missing values. The danger here is failing to select the entire spreadsheet before sorting – always select the square northwest of cell A1 (or convert the spreadsheet to a table). This is where you’ll be glad you numbered each row at the start: to compare with the original.
- If there’s a short note like data source credit that MUST accompany the page and must not get sorted, park it in the column header row to the right of the meaningful headers so it won’t get sorted, lost, or confused with actual data.
- If you use formulas, document the formulas in your Data Definitions page (replace cells with column_names), and copy-paste as value-only as soon as practical.
- Make sure there is only one kind of data in each column: do not mix numeric and text entries. Instead, create extra columns if needed.
- Workbooks should be saved each day of editing with that day’s date (as YYYYMMDD) as part of the filename so you can get back to an older copy. At the end of your session clean up your Changes page, moving To-Do to Done and planning next steps.
Find more spreadsheet guidance here (a set of guidelines recently developed for participants in another project – good links to more resources at its end).
Beyond Workbooks. If you can execute and document your data cleaning workflows in a workbook like Excel, Open Office, or Google Sheets, then you can take your data cleaning to the next level. Knowing steps and sequences appropriate for your specific kinds of datasets will help enormously when you want to convert to using tools such as RapidMiner, R, or Python that can help with some automation and much bigger datasets.
Want more depth? Check out Data Preparation Tips, Tricks, and Tools: An Interview with the Insiders “If you are not good at data preparation, you are NOT a good data scientist…. The validity of any analysis is resting almost completely on the preparation.” – Claudia Perlich
Happy scrubbing! Email or comment with your own favorite tips. Cheers, Ouida Meier