Before this week I was starting to get nervous about the outcome of our project. It seemed like we didn’t have enough data to work with and a limited amount of research questions. When we split into groups according to everyone’s particular angle of interest with our mission statement focus it started to created a better picture.
My group chose to focus on the mission statement’s within the different MSU colleges and map them out in Tableau with their topics. Research question: Are the colleges located North of the river more arts based while South of the river is more STEM based (science, technology, engineering and mathematics)? The tools we chose to use were the Topic Modeling Tool and Tableau Public. We started by creating a Google Doc we shared with each other where we recorded our tasks/steps we take, added all of the colleges and their mission statements and their locations by address.
I was surprised at how quickly our group started putting everything together. So far we’ve met every single one of our tasks we set for ourselves to do, which completely makes or breaks a group project.
We ran into a few problems last week when working with our data and solved them pretty quickly. Since we chose to topic model we began to run the first mission statement through the tool and realized it was too short (one sentence) for the tool to work correctly. We had to find another way to find key words/topics, so we resorted to Voyant.
Although Voyant doesn’t provide us with exact “topics” we are still able to find the key words within each mission statement and use those as topics. Of course the bigger the words in the Voyant graphic, the more times their used in the statement. Here’s an example of the Eli Broad College of Business mission statement after we ran it through Voyant (keeping in mind that we used the stop words list and added a few we agreed on like college, university, michigan, state, and a few others we found within each college).
Trying to get all the words in our word cloud to fit inside the window to screenshot was a problem Laura and I kept running into. We ended up having to refresh the page several times so the words would replace until they all fit. For a few of the colleges we even had to add a few words to the stop words list that we felt were unimportant or less important than the others if refreshing the page wasn’t adjusting them the way we wanted.
Collecting these photos led us to create a Google doc folder. Since we have so many different aspects of our project we realized we needed to keep everything organized and located in the same place. This makes is easy for all of us to access, and easier to refer back to, if needed.
We’re planning on saving each of these word cloud graphics and added them to our map on Tableau Public, which was another interesting tool to experiment with using our data. While Laura and I worked on running the mission statements through Voyant and saving the graphics, Katie worked on the data for Tableau. Dividing up the work in this way allowed us to quickly finish our tasks in order to make more progress on our project. While Katie was modifying our data to work in Tableau she ran into a few problems. We originally thought that by looking up the addresses of each college and finding the latitude and longitude points we’d be able to map then in Tableau. After finding out that Tableau doesn’t accept lat/long points she decided to use complete addresses and map it that way.
I assume that things like this come up a lot in DH projects if you don’t do enough research on the tools you’re using. When we practiced using these tools in class we were always given data, and now we’re creating it and learning more about the format needed for these tools. Sometimes we come to a dead end and have to find a different route/different way to do what we’re trying to do, and that’s DH for you! It’s almost like a maze; there’s only one end and a handful of different ways to get there. If you run into a wall you take another direction and find your way out.
An important part of a DH project is recognizing any limitations and surveying the way your tools work with your data and the data itself. As a class we only obtained the mission statements of the Big Ten universities, but my group was more interested in working with a corpus within MSU so we had to go find our own data.
Obviously, we can see whether or not North of the river is more artsy schools and South is more STEM from just looking at a map but we’re interested in seeing if the mission statements convey the same idea (what are the college’s intentions? What do they focus on? What do they value? etc.) Focusing on one research question we were all interested in is allowing us to put together a project efficiently and easily.
After looking at our project put together as a whole and obtaining comparisons with the key words/topics of each college’s mission statement we’re going to look at MSU’s mission statement and see how the topics we pull from each college compare. Hopefully this will provide a better connection with the other group’s findings as well when we move into collaboration.