How We See Art
An Art Journey Through Centuries
An Art Journey Through Centuries
by Wenge Wang
Have you ever stopped to think about how art changes over time?
The intriguing shift from classical
masterpieces in the 15th century to the avant-garde expressions of pop arts in the 20th century?
Inspired by my personal fascination with
art history, I embarked on a journey through time, and analyzed artworks comprehensively from 15th century
to the 20the century, using the Met (the Metropolitan Museum of Art) API.
May, 2023; Mar, 2024 (Update)
Python, R, Adobe Illustrator, D3.js
Let's start an artful journey with the MET Random Art Odyssey, where artwork data is retrieved directly from the MET API. With a keyword setting to 'painting', the collection emphasizes paintings mainly from the last century. Click the shuffle button to see random artwork!
Check out my code snippet here
I utilized NLTK's part-of-speech tagging and named entity recognition (NER) functionalities to identify and
classify entities such as Organization, Person, and GPE (Geopolitical Entity).
Interestingly,
GPE
mentions and overall categorized topics may be limited in 20th-century artworks compared to other centuries.
This observation suggests a shift in artistic themes.
I further conducted an analysis of word frequency spanning the 15th to 18th centuries to gain detailed
insights
into the
most prevalent topics.
Frequent words like Saint, Child, Virgin,
highlighted above and in the wordcloud, are mostly related to religion, suggesting a significant focus
on religious themes during this period.
Then I conducted another separate analysis of word frequency spanning the 19th to 20th centuries to compare
the
differences between this time period and the previous one.
Frequent words like Woman, Garden, Life, highlighted above and in the wordcloud,
suggest a shift towards themes related to daily life, nature, and human experiences. This shift may
reflect changes in societal values, artistic trends, and the artists' exploration of more
personal and
relatable subjects.
I utilized Colorgram, a Python package, to extract the dominant colors from the overall artworks of each century. The color palettes used in artworks from the 15th century predominantly consisted of earthy tones, reflecting the prevalent artistic styles and pigment materials of the time. Shades of browns and muted tones were commonly observed, indicative of the natural pigments available during that era. In contrast, artworks from the 20th century exhibited a more diverse and vibrant range of colors compared to their 15th-century counterparts.
Click on different centuries to explore ⤵
The Metropolitan Museum of Art (MET) API was utilized as the primary data source for this
project.
The MET API data underwent thorough cleaning and preprocessing for analysis. Artworks from the 15th to the
20th centuries were filtered for focused analysis. Challenges arose when retrieving image URLs due to
restrictions, but alternative methods allowed comprehensive data acquisition from the MET website.
I used D3.js to create visualizations. The Marimekko graph and the two word clouds effectively illustrated the
patterns and transitions found in artwork topics. And the interactive visualization, such as the color wheel,
filtered by centuries, provides users with a visual representation of artists' palettes across different time
periods.
The decision to present color analysis from both RGB and HSL perspectives is based on my
personal experience as an artist, and is aimed to offer a comprehensive understanding of the color palettes
utilized in the artwork.
Through the recent project update, I refined my Python code, enhancing the data cleaning process for more accurate results. By setting clear research questions and narrowing down the time range for my analysis, I found that it brought more focus and depth to my research. This process reminded me of the importance of careful coding and thoughtful planning, which ultimately led to richer insights from the data.