Tag Archives: visualization

Launching Digital Projects from Scratch – Some Advice

Guest post by Adrian Linden-High, PhD student in Classical Studies, Humanities Writ Large research assistant with Digital Scholarship Services at Duke University Libraries in the spring of 2017.

Image tiles of Polygonal Wall: Zach Heater (CC-BY).

Getting a digital project off the ground by yourself can be challenging. With rare exceptions, digital projects rely on collaboration – for the simple reason that it is impossible to unite in one person all the skills needed to deliver a digital product of which fellow scholars will take any note. Still, if you come up with an idea for a digital project, you will almost certainly first have to build prototypes and mock-ups on your own to communicate your vision to potential collaborators. Even this preliminary work in the digital arena represents a challenge for most of us. In this post, I summarize what I have learned from going through the process of starting a digital project from scratch and offer some general advice that will take you to the next, more collaborative, stage of your digital project.

Set realistic goals

Especially if you are new to digital humanities (DH), your project will most likely take much longer to complete than you expect. There are always unforeseen hurdles along the way. Even simple tasks, such as transferring data from one hard drive to another, can be plagued by snags: the data volume in digital projects is often much larger than you encounter in day-to-day computing tasks; if you are working across operating systems (in my case, Mac and PC), you have to make sure you are using a compatible file system for your external hard drives. You get the point. For your project to be a success later, it is shrewd to set realistic goals at the outset. Be ready to scale back your expectations for the initial phase of your project. Just get something up and running. To be sure, this won’t be your final product!

An image stitching project I am working on easily fills 75% of a 1TB external hard drive. I formatted it using the exFAT file system, which is read/write compatible between Windows and OS X and also supports unlimited file size. Click on image for more info on file systems.

Seek out help

Tenaciously seeking out help is vital in the early stages of a digital project. Don’t be satisfied with one answer; get multiple opinions and choose the one that seems the best fit for your abilities, your project goals, and the current phase of your project. It’s worth exploring several tool kits and workflows. There are at least three pools you can dip into for opinions:

  • At your own institution you will most easily have access to students and faculty at your department who are working on digital projects. There may also be a DH unit embedded in your institution’s library (for example, the Digital Scholarship Services department at Duke University Libraries). Beyond this, most colleges and universities have IT professionals who can be extraordinarily helpful, though they may have less time for individual consulting or training (for example, Trinity Technology Services at Duke University).
  • Digital humanities training institutes and conferences are a phenomenal way to connect with people who can help you think critically about your project and connect you with other projects that have a similar vision. Don’t expect to be proficient at any digital technology after only a week or two of training. Most of these skills are acquired and honed over the course of years, not weeks. Premiere digital humanities summer institutes include the Digital Humanities Summer Institute at the University of Victoria, Canada, and Humanities Intensive Learning and Teaching (US). The best outcome you can expect from attending one of these institutes is to have your eyes opened to what is possible and to return home with a welter of new contacts in your address book.
  • Industry is another sphere worth exploring for solutions to your problems. When I was looking for a way to publish and annotate ultra-high resolution visualizations online, I reached out to a company called GIGAmacro that has created such a platform as part of an imaging solution it markets to entomologists, geologists, and manufacturing. I described my project to them and they agreed to let me use their platform free of charge (see GIF demo at the end of this post).

Generate prototypes

It is hard to overstate the value of generating prototypes, especially if you are starting your project as a one-person show and need to drum up interest and funding. Prototyping helps you get a more realistic sense of what is possible and on what time scale. In addition, you will quickly discover how much help you need for a full-scale product and in which areas collaborators are indispensable. Perhaps most importantly, a prototype gives you something to show and share. There is no better way to communicate your idea.

In an initial phase of my digital project focusing on a wall of inscriptions in Delphi, I built an admittedly somewhat clunky mock-up using Lucidchart. Despite many shortcomings, it allowed me to demonstrate what I had in mind and get key advice on how to move forward.

Keep a log

Yes, I know, this one sounds like drudgery, especially for folks from the humanities who aren’t familiar with lab notebooks. But keeping a log for a digital project will save you much time and frustration, believe me. Here are a few reasons why:

  • Tool comparison: When testing tooling options for your project, you will benefit from recording the positives and negatives of each potential digital tool, perhaps even in a table or spreadsheet. It is easy to get confused as to which tool offers which feature. A log also helps you keep track of which ones you have already tried.
  • Reproducibility: Once you have made tooling decisions and are working with the software solution of your choice, you might want to record all the settings you are adjusting in the program. This is especially advisable for things like image processing or stitching where it might take many iterations of fine-tuning settings and rendering to get the desired result. You need an exact record of what you did not only for the sake of convenience and reproducibility, but also methodological transparency. Some programs can generate logs. Take advantage of them!

    If specialized software does not allow you to save a log, consider making screenshots of your tweaked settings for later reference. In this case, I was experimenting with the sliders at the bottom and needed a record of which combinations I had already tried.
  • Project management: A more prosaic reason for keeping a log is that for many of us digital projects are not the main thing we do. They are often side projects and not our bread and butter, at least not yet. And when you haven’t worked on something in a long time, you naturally forget where you left off. A log entry can serve as a great springboard to get back into a slumbering project.
An excerpt from a log I kept while working a visualization of Delphic inscriptions. This entry reflects my struggles with file systems (I had to reformat to exFAT because I was working cross-platform) and preserves some code (a terminal command) I later referred to multiple times.

Start with a small sample of your data

As with any research project, digital or analog, test your idea with a small sample of your data. This is particularly relevant in the digital sphere where you most likely will not have all the necessary skills to tackle all the components of your projects. In the early stages, you can often bridge these gaps temporarily with mock-ups made with less than suitable tools. With a small sample, processing times will be shorter, and data will be more manageable. You are less likely, overall, to get overwhelmed right at the outset. Remember as well that you are testing whether a process will work or not. If you try to be as comprehensive as possible with the first iteration, you run the risk of wasting time and resources on a flawed approach. You can always scale up your project once something small works well. Starting with a small sample, finally, is a great reality check in terms of your hardware capabilities. If your current hardware configuration struggles to process a small sample, you know you will need to find more powerful computers to tackle the next stage.

The animated gif above illustrates how I progressively added more tiles to an image stitching task, going from three to ca. fifty. My laptop crashed when I attempted to stitch 100 tiles, whereupon I had to migrate to a lab with beefier machines. There I finally succeeded in stitching my entire batch of 600 image tiles.

Prioritize project components

Once you have determined the building blocks your project falls into, it is time to decide what to tackle first. Many variables enter this equation and they will be weighted differently from project to project, so it is hard to give general advice. One thing worth thinking about is what potential collaborators and funders are interested in. More likely than not, they will not be looking for a flashy demo, but solid ground in terms of planning and project feasibility. Some Digital Humanities grant programs focus on social infrastructure, planning, and collaboration as much as (or more than) technology per se. Having said that, a prototype of your project can be a phenomenally effective way to communicate your idea. If it is visually appealing, all the better. Above all, as I stated at the outset, remember to be realistic: you have time and resource constraints, and everything will take longer than you expect. Projects can get derailed by focusing too much on creating something flashy to the exclusion of more essential project components, like content creation, target audience assessment, data structuring, or rights/fair use assessment, to name a few.

A more advanced prototype of the visualization of the wall of inscriptions in Delphi using the GIGAmacro Viewer which includes features such as support for extreme resolution images, polygon annotations, and many more. Click on the image to have a look yourself!

 

Good luck!

For a project management perspective on developing a digital humanities project see this excellent overview published by the Emory Center for Digital Scholarship. Another valuable resource is the Development for the Digital Humanities website.

 


Adrian Linden-High is a sixth-year Classical Studies PhD student at Duke University whose research centers on inscriptions and papyri that capture what everyday life was like in the ancient world. Currently, he is working on a rich archive of inscriptions from Delphi recording more than 1,000 slave manumissions (see digital visualization prototype). In the spring of 2017, he served as a Humanities Writ Large research assistant with Digital Scholarship Services at Duke University Libraries.