My co-faculty and I took the data visualization course we taught at the University of Washington. The bag consists of a rather diverse collection that I have placed myself in a set of treks across various areas of California.
Our students are fairly accustomed to the quirky, hands-on activities we ask them to attend most classes, but it seems a bit like us.
In this article, I will focus on the following two points, which together describe the importance of domain-specific integration into data science education:
- Our description of the actual tasks we perform on these rocks.
- Deeper in the subsequent discussions – focusing mainly on the point where they do so and the deeper connection to data science.
What to do if a bunch of rocks?
Once students sit in their respective groups, we ask them to do the following:
- Two rocks are selected for each group.
- Try to formally identify the rock without any internet or mobile apps. At this point, most students can be sure that the rock appears to be igneous, sedimentary or metamorphic.
- Refine their initial guess by now using their electronic resources. Now, students become more specific and can identify Scoria, Slate, Red Jasper, Gneiss and many other rocks in the series.
- Design and implement a chart (using software or on paper) to compare its rock mass or display fascinating information about one of them. They are encouraged to search online for support data such as hardness, mineral makeup, potential uses, and more.
Once finished, they submitted the visualization to us and we had a class discussion.
What is the relationship between rock and data science?
A lot has happened.
As we walked around, the students shared many insights about their various rocks. In many cases, the discussion focuses on the practicality of the specific visual approach taken by students.
For example, a group chooses to compare their two rocks through a data table containing various relevant information points. This led to discussions about data tables yes In fact, a data visualization is particularly useful in two situations:
- When your data is limited
- When a user is able to select precise data for their purpose, it is important to
Other conversations revolve around the effectiveness of the region as a coding, the particularity of the color scale, etc. Discussion on all standards for the data visualization course.
Once we finished the initial conversation, I asked a more class question:
“So far, we’ve talked about the standard visual elements of the chart. We could have discussed these with any type of data. So why have you had trouble bringing a large piece of rock to the class and asking you to identify them? What’s the point?”
The class stared blankly. That moment dragged on. Then, a student raised his hand hesitantly.
“Well…so can we be comfortable working with strange domain names or something like that?”
precisely! We rarely mentioned this to students before, but this activity did bring the focus home. As the ultimate designer and engineer, working on data visualization, and more broadly, in data science, for these students, it is necessary to know how to work with areas they may not be familiar with.
If you are reading this, that’s true for you. As a data expert on the team, you seldom become a domain expert, and you have to adapt to the data you give you. Sometimes it’s very fast.
In the previous article, “Three Foundations of Data Science,” I entered this in more detail. The first two building blocks (statistics and computer science) are very important. That is, actually data From the domain. Without domains, there is no need to do data science.
As a data scientist, while you will be supported by domain experts, you still need to design solutions and write code that corresponds to data you may be very unfamiliar with. Therefore, it is very important to get exposure of this reality as part of data science education.
My co-teachers and I teach in the design and engineering department, and students are largely interested in pursuing areas such as UI/UX research and data engineering. We chose to make them work with rocks because we know they are unlikely to know too much about them in advance (at least at the level of detail required).
The lack of prior knowledge makes everything different.
The final thought
If you are reading this article, I guess you are training to become a data scientist or are interested in doing so. Maybe you are already alone and just sorting out your knowledge.
Regardless of your location, my opinion remains the same: Every opportunity you get is exposed to new data. Essentially, literally, every field, every discipline, every topic known to humans has some kind of data, and a group of people who are interested in getting insights about it.
You may be the one they seek help.
