X.6: Choosing Your Data Katana
Guest post from Matt Moellering
After a long day managing the S3 shop, your Operations Officer walks in and drops a bombshell: “We’re going to become a data-driven organization, and you’re going to lead the charge.” You think to yourself, “Fantastic—just one more thing to add to the to-do list right after world peace.” But if you’re reading Downrange Data, you already know that transforming your unit into a data-driven organization isn’t just another buzzword bingo square to check off. The following guide will help you select the right data tools to cut through the complexity.
Principles
There are endless ways to "use data," and the age-old question of what "data literacy" really means has kept philosophers busy since the dawn of Excel. But when it comes down to it, the golden rule, especially in industry, is: "How will this generate value?" For the Department of Defense (DoD), this means accomplishing the mission—whether that’s outmaneuvering the enemy or just outsmarting the PowerPoint slides. As Erik Davis succinctly puts it, “cheat with data.” While there are numerous ways to "cheat" with data, any method you choose should focus on advancing your organization’s objectives and enhancing the OODA loop of your core tasks.
Analysis Tools
There are more data analytics tools out there than there are acronyms in the military, but for the sake of simplicity (and sanity), we’ll focus on tools already at your disposal within the DoD. We’ll break these down into three categories: Microsoft Office tools, programming languages/tools available on most government networks, and Software as a Service (SaaS) tools that your unit is probably already paying for (because who doesn’t love a sunk cost?).
Microsoft Excel:
Ah, Excel—the Swiss Army knife of data tools. Whether you’re a data whiz or just trying to make sense of that mess of numbers the LT sent over, Excel’s got your back. With tables, charts, and formulas, it lets you organize, analyze, and visualize data in ways that would make your high school math teacher proud. For those who are in the middle of the bell curve, don’t forget that most of Wall Street still runs on Excel and, unlike some of the tools previously mentioned, everyone knows what you are showing them.1
Pros: Excel is like the universal remote of data tools—widely known and incredibly flexible. Plus, some of the world’s top data analysts still swear by it, as evidenced by the Excel World Championships (yes, that’s a thing).
Cons: Automation in Excel is a bit like trying to herd cats, and it doesn’t play well with large datasets. If you’re looking for advanced data processing, you might want to look elsewhere.
Microsoft Power BI:
This business analytics tool from Microsoft allows you to create interactive reports and dashboards, making it easier to turn data into insights and, hopefully, make sense of that never-ending stream of information.
Pros: Power BI is available on government computers, so no need to deal with the S6 telling you you can’t use it. It lets you create data dashboards and even post them on the DODIN.2 It also supports data automation—assuming you can get your hands on the data.
Cons: Automation can fail, and when it does, you might find yourself on a first-name basis with Murphy’s Law. Always have a PACE plan—“Primary, Alternate, Contingency, and Emergency”—and remember, “Two is one, and one is none.” It’s also a time-intensive initiative, so consider delegating it to an ORSA or, if you don’t have one, a technically solid junior officer / NCO. Here's a suggestion – find the talented teammate who is currently stuck in a staff job but still owes a year or so before they can leave the service. This is a fantastic task for them to work on to level-up their resume and automate tasks for their organization at the same time.
R/Python
R:
R is a programming language designed for statistical computing and data analysis. It offers a wide range of techniques, from linear modeling to time-series analysis, all while making your data look like a work of art. There are several offerings of Rstudio, an integrated development environment that allows you to write code and manipulate data securely on a government computer. Additionally, most of the programs that have R studio usually have the option to host R Shiny applications which allow anyone to view your data solutions.
Python:
Python is one of the most popular programming languages in the world and is really strong at both processing data and using machine learning. It is also more flexible than R, and, just like R, is available on most government platforms. There are multiple ways to use Python, but one of my favorites is a Jupyter Notebook. Jupyter Notebook is an open-source Python web application that allows you to create and share documents containing live code, equations, visualizations, and narrative text. It’s like a digital notebook for the data-obsessed, supporting various python libraries and used in everything from scientific research to machine learning. Just like R shiny, there are several different python native web applications (Streamlit, Dash, Shiny, and Flask – just to name a few) that allow you to easily deploy applications.
Pros: If you can code, these tools are like giving a kid the keys to the candy store—highly flexible and capable of creating beautiful visualizations. They’re also accessible through most government systems, so no need to ride dirty trying to build it at home.
Cons: Like Power BI, these tools can be a bit temperamental when it comes to automation and require some coding expertise. Bring in your data science team to help smooth out the rough edges and ensure everything aligns with your endgame. Also consider seeking support from a higher headquarters data team, the software factory or the AI2C for further assistance.3
Data Platforms: Palantir/Advana
Data Platforms:
The DoD has invested in several advanced programs, such as Advana and Palantir, developed by leading contractors. These tools are designed to integrate, analyze, and visualize large datasets from multiple sources. Given that these programs are already established, I recommend prioritizing their use. However, it’s important to note that skill levels among users can vary, and many units may still be unfamiliar with the full capabilities of these tools.
Pros: Using data platforms that the Army is already paying for is a no-brainer.
Cons: Before you dive in, assess the quality of the data in these systems. Start by conducting an audit—because the last thing you need is a shiny new platform full of garbage data. Let’s face it, most units have garbage data.
The Raw Material: Data Itself
Speaking of which, your data is probably in rough shape. If you plan to use data from Army systems, make sure to spend significant time validating it with subordinate units and retrieving it from existing Army systems. Remember that higher commands and HQs rely on data from subordinate units to populate their dashboards, which then inform their decisions. Ensuring data quality is crucial because even the best data tools won’t be effective if they are working with poor-quality data.
The Way: DevOps and Kanban
DevOps:
DevOps is like the operational art of software development—a set of practices that combines development and IT operations to enhance collaboration, automation, and continuous delivery. It’s all about streamlining the software development lifecycle to ensure that your applications are deployed efficiently and reliably. Just as JSOC enhanced its effectiveness by integrating intelligence with operations, the software community developed DevOps to ensure that those writing code are closely aligned with IT professionals who implement it.4
For the more advanced units, consider DevOps as a way to develop and deploy applications with military precision. While some Army units are already equipped for this, others might still be in boot camp. But don’t worry—Agile methods can help whip your unit into shape. Agile is like a series of tactical drills in the field: it focuses on iterative progress, frequent reassessment, and continuous improvement. Just as military units adjust their strategies based on real-time feedback and evolving conditions, Agile teams regularly review their work and adapt their approach to better meet objectives. This helps teams stay flexible, respond swiftly to changes, and deliver results more effectively.
Kanban Boards:
Kanban boards are visual tools used to track and manage workflow through columns representing different stages of work. They’re great for visualizing tasks, managing priorities, and keeping the chaos at bay by limiting work in progress.
I’m a big fan of Kanban boards for managing operations—they’re like turning a to-do list into a well-oiled machine. They help you prioritize tasks and automate delegation, transforming chores into data-driven decisions. The Army already provides Microsoft Planner, which is a great tool for this purpose. Additionally, there are better kanban boards such as Jira and Gitlab issues that your unit may already be paying for. If you happen to be a fan of Notion, you will appreciate tracking tools like this.
Conclusion
The key to successful data science is creating value for the end user. Organizing data is only useful if it doesn’t turn into a time sink—like those never-ending PowerPoint presentations we all love to hate. Following the Downrange Data philosophy, this article makes no mention of PowerPoint because it's not a data tool. Focus on building systems that reduce the need for pitching and enhance the core aspects of your warfighting function. Automate where possible, and continually integrate more data feeds to keep your operations sharp and your presentations mercifully brief. Remember, just as every Samurai carried two swords, you should equip yourself with multiple data tools.
CPT Matthew Moellering previously served at AI2C after completing the AI Scholar Program. Matt also volunteers for the Irregular Warfare Initiative serving as a co-host for the Irregular Warfare Podcast. Matt is currently an Army University Press Dubik Writing Fellow.
Department of Defense Information Network (DODIN) if you are on a computer provided by the military you are probably on the DODIN.
Army Artificial Intelligence Integration Center. Home of one of the three Army Software Factories.
Joint Special Operations Command. A joint component command of USSOCOM charged with studying special operations requirements and techniques to ensure interoperability and equipment standardization, to plan and conduct special operations exercises and training. The Joint Special Operations Command also oversees the Special Mission Units of U.S. Special Operations Command.
 |
|