Translation Device

    For part 1 of Translation Device, I am hoping to investigate the collection data of Cleveland Museum of Fine Arts. My initial hope is to work with the data of a contemporary art collection at a major public museum. After multiple rounds of research, the only museum I found that has a functioning API and met my criterion with its collection is the Cleveland Museum of Art. With the collection data, I am specifically interested in the time of accession/acquisition because the acquisition of an artist's work by a major public museum is generally seen as a landmark in their career, and that they have joined the rank of recognized artists. The key questions I am interested in asking are: on average, at what age does an artist’s work get acquired by a major public museum’s collection? How does their gender, ethnicity and nationality affect that number?

    Acquisition is this kind of landmark event in an artist’s career. If your work got acquired by a major art museum, it basically just means you’ve made it. You have joined the rank of established artist. I am interested in looking at the question: at what age does an artist’s work get acquired by a museum? And does factors like gender affect that? For example, if you are a female artist, would your work get acquired at a later age/stage of your career than a male artist?

    I ended up working with the data of Cleveland Museum of Art because a lot of museums with a major contemporary art collection do not have a functioning API, or even a data dump, which is very unfortunate. The Cleveland Museum has around 2000 modern/contemporary artists in their collection, which is a pretty sizable collection of data. so I thought that this dataset still would be able to tell us some interesting trends in the general practice of museum acquisition.

    So for the project, to find out the range of artists ages, I need to parse through a lot of data and do a series of calculation to figure out the numbers. Because the data I am looking for is not just there on their database. So I actually wrote a program in Python to parse through that, just because I am unsure if p5 would be best tool to do all those relatively complicated parsing.

    The outcome of the data parsing was very interesting because it turned out that female artists actually got acquired at a slightly younger age than male artists. To fully understand why this could be the case, I also did a bunch of smaller histogram tests at the end of my python code. There is no jump in the number of young female artists acquired by the museum at any particular year/time period, but there has been a slow and steady increase in the amount of female artists acquired in every age group and slightly more with the younger groups. For both male and female groups, there has been an overall slow shift towards the older age groups from the younger age groups. Since there has been no outstanding point over the years where there is a drastic change in the trends of the data, I decide to simply focus on my initial question about the average age of the artists at times of acquisition. There is no particular critique here but I still find these results interesting.

    You can check out my Python code here.

    Below are some of the initial sketches I have for my data visualization:

    I chose to go with the third sketch because while my graph mainly tries to represent the age distribution of male artists versus female artists, and that the average female artists’ age at times of acquisition is actually lower than male’s, I want to emphasize that the female artists still only represent about a fourth of the entire collection of artists. With that in mind, I decided to use this nested structure, where the graph for female artists is placed within the total number of artists, and respectively with the graph for male artists as well.

    I am very proud of the end results. To achieve a less jagged effect of the shape, since there could be a drastic difference from the number of artists acquired from one year to another, I used a moving average method to calculate the data. Just like how many COVID tracking website represents its data using a 7-day average, I calculated my data using a 7 point average as well. I decided that I would compromise the representation of the precise number of artists acquired every year because I am more interested in the general trends in terms of age group.

    I think my final design successfully answer two principles in data: examine power and elevate emotions and embodiment.