From Meghan: This is a guest blog post by ecologists Isla Myers-Smith and Gergana Daskalova from the University of Edinburgh. I loved their comment on my post on our new lab notebook backup system and asked them if they could turn it into a guest post. I was very happy that they agreed! Isla and Gergana are off to the Arctic this summer with the Team Shrub field crew for another year of hopefully successful digital data collection. To find out more about their research check out the Team Shrub website and blog (https://teamshrub.com/).
Two things have really changed my academic life over the past five years: the first is embracing GitHub for version control of code, data, manuscripts and my research group’s individual and combined science, and the other is switching over to digital data collection. For ecologists who haven’t made the switch from paper field books to iPads and digital data collection it is not as scary as you might think!!!
Caption: Collecting plant phenology data – the recorder sitting in the back with an iPad! (photo credit: Jeff Kerby)
The benefits of going digital
Digital data collection can be more rigorous with error checking as data are collected to prevent mistakes. Data can be better backed up. And finally, it forces us to put thought into the structure of data before we collect it (significant digits, continuous or categorical data, are the data unrestricted or constrained to a particular range or particular set of values, etc.), which helps down the road when it comes time for analysis. Digital data collection has saved days, if not months, of data entry each year for my team and has allowed us to go from ecological monitoring in the field to analysis of results within hours instead of days. Our work flows are streamlined and our iPads are waterproof, so data collection can occur under any conditions – and we work in the Arctic, so we experience it all from wet to dry, hot to cold, rain, snow, you name it.
In 2015, my research group moved to using iPads for data collection. At first I was skeptical. I had used electronic devices in the past with mixed success and had given up, but of course technology had progressed and working with multiple devices and syncing data is much easier now. So, my PhD student and postdoc convinced me to lose the field note books (we still carry them, but now they are for our personal notes rather than the main data collection efforts) and go digital.
Caption: Shenanigans captured in the field by iPads (notice the iPad in the top right). Don’t worry, no principal investigators, PhD students or field assistants were hurt in the making of this blog post. That “invigorating elbow drop” in the middle of a very tedious field protocol was all fun and games. (photo credit: Joe Boyle)
Our digital data collection workflow
Each member of the field team has access to an iPad for data collection. We use the Numbers app or the Sheets app on iPads to enter data (thus skipping the paper and pencil field book stage). Before heading out to the field, we prepare our spreadsheets, which also helps us to think about what data we should be collecting specifically and what format is best for collecting those data. We have waterproof cases for the iPads, so we can use them under any conditions including quite heavy rain. Entering the data straight into a digital format helps avoid transcribing errors when copying information over from field books to spreadsheets, and it means we can go through lengthy and detailed field protocols much more rapidly – saving time in the field for more data collection (or a bit of fun!).
Caption: Our digital data collection workflow.
Practical benefits of digital data collection
For example, for our long-term vegetation monitoring, instead of having to manually write down each species name and how tall the individual is (for 1200+ records), we have a drop down menu with species names, ordered by commonness, and when the observer says “Salix arctica, 42.7, three leaves“, the recorder can select the species from the drop down list, note the individual’s height numerically using the appropriate decimal places and then increment a count for the number of leaves sampled. This speeds up the data collection in the field and removes the need for data entry afterwards.
Caption: Sometimes iPads lead to innovative digital data collection such as this hand-based numbering system for field plots that worked great, until I accidentally used the sports mode on the iPad camera and produced all of these lovely photos. (photo credit: Isla Myers-Smith)
Digital data collection in the field also shortens the amount of time you have to wait till you start unwrapping your data presents! We don’t have to spend weeks digitizing our data from field books and we can now load them right into R and analyze them when we get back from a day in the field. Making a few quick plots of the raw data is also good for spotting potential errors, but also adapting your field data collection protocols to make sure all the critical data are collected.
Caption: A field data present compiled a day or two after data collection from data collected on an iPad in 2017 resulting from the International Tundra Experiment species pool protocol. You can read all about these exciting results in our blog posts on Mark Vellend’s ‘The Theory of Ecological Communities’ or hopefully in a paper someday soon!
In the field, we take photos of our experimental plots, species that we can’t ID, and the field team in action that can be quickly associated with the data themselves. We also take regular screen captures as we are working, providing a digital backup during data collection. At the end of the day, we sync our iPads with each other and with multiple other devices including computers and hard drives (and that is without internet in our Arctic field sites, this an even quicker process if you have access to the internet) and as soon as we are back to the internet world, we sync our iPads to the cloud. We use version control when appropriate including version numbers in file names. Thus, all our data are backed up and synced, but also integrated into our data handling workflows, so that we can for example save files as .csv and import into R for incorporation into the master dataset and proceed with data checking and analysis.
Caption: Screen shots are a great way to create a digital back up of data collection when out in the field. Fujitsu became our word of the summer in 2016 (it was an auto correct from another word we were using). If you write out Fujitsu or any other word and then delete it again, that becomes your last entry after a bout of data collection. So, if you are accidentally shaking your iPad as you move to the next plot, you won’t delete your last data point. Alternatively, you can turn off the shake setting for delete, but it is kind of useful for rapid data collection.
For our research group, Team Shrub (https://teamshrub.com/), combining these digital data collection practices with version control using GitHub dramatically increases our efficiency and data security when we can carry out laborious field monitoring protocols and update analyses with new data. Of course, we quickly spend all of the time that we have saved on collecting new types of data such as data from drones, trying to improve our data analysis skills, building extra complex models or creating even larger datasets, but I think the result is more rigorous and exciting science.
Captions: Top) Measuring plant height out in the tundra; Bottom) A tundra nap after collecting thousands of rows of data on plant species composition and traits – notice the iPad! (photo credits: Gergana Daskalova)
So, if you haven’t tried completely digital data collection, I say take the plunge and give it a go! Four years and counting, and I haven’t looked back yet. And, if you want to find out more about GitHub for research groups check out our Coding Club tutorial on the subject!