The Green Edge ice camp data paper

# The Green Edge ice camp data paper
## 📝<br/>Green Edge legacy meeting (Nice, France)
### Philippe Massicotte et al.
### 2019-11-06

---

Yes, I am that person that sent you all these emails concerning the ice camp data paper. **By the way, thank you to those who answered.**

---
class: my-one-page-font, inverse, center, middle

# Working with data from others

---

# Why is this tedious?

--
  
- Many sources of data.
  
--
  
- Different formats (xls, csv, mat, dat, etc.).
  
--
  
- Acquired at different temporal/spatial scales.

- Metadata rarely available.
  
--
  
- **Data structure highly variable among the researchers.**

---

# The consequences of working with non-structured data

> Most data scientists spend only 20% of their time on actual data analysis and 80% of their time finding, cleaning, and reorganizing huge amounts of data, which is an inefficient data strategy.

- Uniformizing data is time-consuming *(unfortunately not often recognized)*.

- This is why we have worked hard to uniformize the Green Edge data in order to **publish it in a data paper**.

---

# What is a data paper?

*A peer-reviewed document describing a dataset, published in a peer-reviewed journal.*

---

# Main goals of a *(the ice camp)* data paper

1. **Compile** and **structure** the data to make it **easily reusable** by other researchers.

2. Make the data **discoverable** for the community.

3. Make the data **citable**.

---

# The Green Edge ice camp data

A tale by numbers: how to navigate *(almost)* blindly in the data of others.

---

# The data: a few statistics
  
--

- 3 000+ raw data files that have been regrouped into `\(\approx\)` 300 CSV files.

- 5 000+ lines of code just to *tidy* the data.

- **25 000 000+ observations!**

---

# Data availability

The data was deposed online [Sea scientific open data publication (SEANOE)](https://www.seanoe.org/). DOI: 10.17882/59892

---

# The Green Edge data papers

*I wish I had a great subtitle to write here...*

---

# Two data papers in preparation

1. Ice camp 2015-2016: Philippe Massicotte

2. Amundsen 2016: Flavienne Bruyant

---

# The Green Edge ice camp data paper, it is...

- 109 authors (≈12-15 commented on the paper, thank you!)

- 43 affiliations/institutions

- **Too many emails to figure out the data!**

---

# Many institutions involved

---

# Data papers

Submitted to **Earth System Science Data (ESSD)**

> Earth System Science Data (ESSD) is an international, interdisciplinary journal for the publication of articles on original research data (sets), furthering the reuse of high-quality data of benefit to Earth system sciences.

Impact factor: **10.951**

Impact factor (5-years): **9.899**

---

# The structure of ice camp data paper

Divided into sections to present:

1. Physical data

2. Underwater bio-optical data

3. Nutrients

4. Bacteria and Phytoplankton

5. Zooplankton

---

# One more thing...

<SPAN STYLE="color: #FCD116; font-size: 50pt";><b>The paper has been accepted!</b></SPAN><br><br>

---

# The reviews are very positive

> The quality checks are appropriate and the process of reviewing the data is up-to-date and grants for **usefulness of the data** to other potential scientists.

> **The presentation is of high quality** and I don’t see any inconsistencies that could raise suspects that the data are erroneous.

> ... the data presented hold potential for being reused in the future for comparison and further elaboration.

> The authors produced **an impressive, integrated data set**.

---
class: my-one-page-font, inverse, center, middle

# Lessons learned and recommendations

After all, it would not be fun!

---

# Lessons learned (1/2)

Although initial recommendations on good practices about data collection were communicated to all scientists, **extensive efforts had to be performed to assemble the data**.

Thanks to Marie-Pier Amyot!

---

# Lessons learned (2/2)

To reduce possible errors and increase the overall efficiency of science:

1. **A uniformed data management plan** should be prepared and distributed (**and respected!**) prior to each mission.

2. Dedicated data management specialists should be involved from the beginning of the project.

---

# Basic recommendations for data management (1/2)

- Whenever possible, **do not use proprietary file formats** such as Matlab or Excel:

- Use plain text files (ex.: CSV, TXT or TAB)

---

# Basic recommendations for data management (2/2)

If you use a spreadsheet program (ex.: Excel), keep your data arranged as rectangular tables. Otherwise, **it makes data importation difficult.**

.pull-left[
<center><img src="https://exceljet.net/sites/default/files/styles/function_screen/public/images/formulas/join%20tables%20with%20INDEX%20and%20MATCH%202.png?itok=HK29obOJ" class="centerImage" height="300"></center>
]

.pull-right[
<center><img src="https://www.excel-easy.com/examples/images/split/split-worksheet.png" class="centerImage" height="300"></center>
]

---

<br>
<SPAN STYLE="color: #B2CCE5; font-size: 45pt";>Many thanks to those who helped directly or indirectly with the data!</SPAN>