By: Vidushy Avasthi, Helena Miranda, Sophie Diebold

WEBSITE: https://vidushyavasthi.wixsite.com/info-419-term-projec 

OBJECTIVES 

This project reviews public art in Metro Vancouver through Information Visualization. It looks at how the types of art, occurrence, and ownership of public art have changed between 1901 and today, and places public art trends in a historical context.

 Public art is defined as artwork that is open to the general public, regardless of whether it is situated on public or private property, or if it has been purchased with public or private funds (Tate). This topic is relevant due to the institutionalization of art and the controversial opinion that art is only art when in an institution, often one that requires a fee to gain admission and is therefore inaccessible to much of the general public. Accessibility is an interesting facet of public art since Vancouver is a relatively young city where the lack of affordable space is a major problem. There is a rapidly rising industrial and commercial land value that as a consequence promotes the loss of affordable spaces where artists can share their work (Vancouver City Council, 2019, p. 1). 

Our objectives for this project were to visualize the prevalence of public art according to type of art, neighborhood, year of installation, and whether they are owned by public or private bodies. We wanted the visualizations to show a narrative where viewers can see how attitudes to, importance given, funding given to public art projects have evolved over the past century. The InfoVis show how art popularity changes by neighborhood, and how ownership of public art influences the area where it is located.

Intended users are local and visiting artists trying to understand public art trends in Vancouver over time, space, and according to type of art (sculpture, media work, mosaic, relief, among others), and tourists or anyone else wanting to explore and learn more about the artistic landscape of the city.

 

DETAILS ABOUT THE DATASET

Some questions we wanted our InfoVis to answer were:

• How does the popularity of a type of art change throughout the years?
• Are more pieces of public art privately or publicly owned?
• Do certain neighborhoods have more privately owned art? Why or why not?

The data used for this project was sourced from the City of Vancouver Open Data Portal (https://opendata.vancouver.ca/explore/dataset/public-art/information/). The portal has hundreds of datasets that are updated daily, all of which are specific to the City of Vancouver. The dataset we have chosen contains the following attributes: Title of work, artist project statement, type of art, status (whether it is still in place or not), site name, site address, primary material, URL to artwork, URL to an image of the artwork, ownership, neighborhood, on-site location specifications, coordinates, geographical area, description of work, year of installation, and photo credits. The data we sourced needed to be cleaned to create visualizations. We used Tableau Prep to polish the data. This process involved eliminating repeated values and identifying the attributes we would need to work with to fulfill our objectives. This cleaned data was then imported into Tableau Desktop to create visualizations. We settled on the idioms, color schemes, and fonts beforehand to allow all visualizations to look cohesive. These tools were familiar to us through coursework hence they were not too difficult to use for basic visualizations, but since we hadn’t yet explored the full potential of either Tableau Prep or Desktop, going through multiple features and identifying the ones we needed was an overwhelming task. Tableau’s self-contained help videos and discussion forums were very helpful for when we got stuck.

INFOVIS

Both the website and the InfoVis were created with this color palette in mind for a more cohesive visual structure. 

1 – TRENDS IN TYPE OF ART INFOVIS 

The dataset includes an attribute called ‘type of art’ and we thought it could be interesting to explore how these types of art in public art changed across the years since 1901. The dataset also contains the status of these public art, that is, if they are still in place or not, or if they have been deaccessioned (removed for selling or discard). We thought including their status would help the understanding of trends in public art across time since we are able to see which type of art ‘survives’ or not different periods of time and thus how mediums, materials, cultural and social aspects that are associated with these different types of art change throughout the years (e.g. new media work with LED lights versus classical bronze busts of historical figures).

The categories in the dataset were mixed and sometimes overlapped. For example ‘Memorial or monument’ and ‘Figurative’ sometimes incorporated the same type of art so I created a definition for each category and re-organized the dataset according to these new definitions. The definitions and examples can be found on the website. 

We tried many ways of encoding the data like with the treemap and packed bubble chart on Tableau Desktop but the bar chart worked best alongside the filtering options of year and status. Moreover, the line chart helps having a general and less detailed overview of the different trends of types of art throughout the years, something the bar chart does not allow. In this sense, Tableau was really versatile and thus we were able to quickly create many different InfoVis with the same attributes to see which ones communicate more clearly the trends of type of art throughout the years. 

From the beginning, we thought bard charts would be the ideal form of encoding a number of types of art throughout the years, although we did some other experiments mentioned above. 

This was one of the first sketches we did when brainstorming the InfoVis for our project. It was helpful to see what I need to do with my data right away (creating a calculating field to find the number of each type of art and also getting the types of art themselves). According to the principle of expressiveness and effectiveness, the categorical attribute of type of art was encoded with identity channels of spatial regions and hue were assigned to each of them so it is visually more interesting and cohesive with the theme of public art. The attribute of count of type of art was encoded with the magnitude channel of length in a common scale, especially because it is cumulative, that is, indicates a counting number, so length felt more appropriate than spatial position. Moreover, to avoid lying charts, the count of the type of art axis’s numbers are fixed when the viewer filters the data, so they have a fair comparison of the count of type of art throughout the years. 

The cons of this Infovis is that it should include more images and perhaps the titles of the artworks, but there were enough attributes included in the visualization already and it would make it overwhelming to look at. That is why we included some examples in the website, so the viewer sees what we mean by each type of art and also to highlight some of the iconic public art of the city of Vancouver.

Some conclusions can be made about the trends in type of art, like the fact that sculpture is the most prevalent type of public art across time and one of the most lasting types of art as well. In 2016 we see an explosion in the number of Murals, adding up to 42 murals, which is the highest rate we see within a year compared to the other years and art types. The second biggest rate of type of art in a year was Two-Dimensional Work in 2020, adding up to 27 artworks. In the beginning, starting from 1901, the prevalent Types of Art however are Memorial or Monument, Totem Poles, Fountain or water feature, and Figurative up to the 1950s. Then we see the emergence of Sculpture as one of the main types of art, followed by Site-Integrated work around the 70s and 90s. Media Work, Murals, Two-Dimensional Art, and Site-integrated works gained force around the 2000s and this pattern continues until 2024 although murals seem less popular since the pandemic, after 2019.

2 – INSTALLATION PER YEAR INFOVIS 

This is one of the simpler charts in our project. The goal here was to visualize any notable trends in the number of installations per year from 1901 (first public art installation according to our dataset) to 2024 (the present day). The main attribute we looked at to plot this visualization was the year of installation. Tableau Prep read this attribute as an integer instead of a year (so a comma was included after the thousand), so we had to make sure the program read it as a year. It is interesting to note that there is some correlation with our “Trends in Types of Art” InfoVis. Evidently, 2016 was a good year for murals.

This is an interesting visualization when taking into consideration the fact that Vancouver is a relatively new city. This proves that over the years, there has been a push for public funding in favor of those things that help foster culture. Art has the ability to raise awareness and educate people, and when it is displayed publicly, these statements make their way to more audiences. It would be interesting to see a data set that approaches public art in Vancouver through demographics to note what kind of voices are being most amplified. 

 

3 – LOCATION MAP INFOVIS 

The idea for the location map was to have an interactive viz that would allow users to navigate across the city and see neighborhoods where art installations are located. The use of a GeoJson file type made it easy to create a map since that file comes with the geospatial component integrated with the file. We used the coordinates to mark places on the map where art installations are located. These marks were then color coordinated by neighborhood to make them more distinct from other marks that appeared nearby but were in a different neighborhood. To make this map more visually interesting, instead of using regular Tableau-provided shapes for marks, we imported a vector icon of the Mona Lisa. We used that as a mark instead after adjusting the size for legibility. Then we added a hover interaction that will allow users to highlight all installations in a single neighborhood. This action can also be done by using the filters on the right-hand bar.

A positive about this InfoVis would be that it’s interactive and easy to use, however, critically speaking this InfoVis would have benefited from the addition of filters to make it easier to distinguish between types of installations between neighborhoods, instead of just having their locations mapped.

Sketch:

4 – OWNERSHIP DISTRIBUTION INFOVIS 

The ownership attribute proved that the dataset needed some cleaning in order to produce a visualization that would be effective and easy to digest. With over 40 different values, each being quite specific, the visualization could probably only be read by someone who is familiar with the city of Vancouver and the various organizations and companies that occupy it. Below a screenshot of what this looks like in Tableau prep:

We encountered many vague labels such as “university” and “Bosa”. Although not ideal, we had to make assumptions about these. We assume that “university” could possibly represent UBC. When Googling “Bosa”, we encountered “Bosa Properties” and “Bosa Foods”. Both would be in the private category, so we are not putting too much importance on what it represents. After dissecting the values within this attribute, we used the grouping feature in Tableau Prep to form the following categorizations:

• Public
• Private (Owned by a person)
• Private (Corporation)
• Private (Non-profit/Society)
• Private (Institution)
• Privately owned (no further information)
• Unknown

After cleaning the data through Tableau Prep, I exported it to Tableau to start working on the visualization. I realized the data somehow got mixed up in the cleaning process, and the program was reading the “Ownership” attribute as one of the other attributes. I tried cleaning it again and the same thing kept happening, so I had to create a new dataset that only included the “Ownership” attribute. I ended up using this to plot the pie chart in Tableau. As stated previously, we followed a specific color palette, which was implemented in this visualization as well. Each distinct color represents an ownership category. The pie chart is a good way to visualize the distribution of ownership’s because it divides the big picture to show parts of a whole. It is easy to see what percentage of public art is owned by who. 

It’s nice to see that most public art in Vancouver is actually Public! The runner up was “Privately owned” with no further information. Because the dataset did not provide any more detail, there is a knowledge gap present, and all pieces that fall within this value could, in actuality, belong to any of the subsections we came up with. The third largest category were those with unknown ownership, so we have another knowledge gap as well. From the data we had access to, most privately owned pieces are owned by a non-profit organization, or a society. It was also surprising to see that two pieces in the dataset were owned by a sole person. One of them being owned by Bob Rennie, an art collector, and another one being owned by the artist Janet Echelman. Overall, the method in which the values were categorized is somewhat reductive, but it shows a general idea of how ownership of public art is distributed within Vancouver. This process had to be done to make the understanding of the data more accessible. 

Bibliography
Munroe, D. C. (1973). Public art in Vancouver (T). University of British Columbia. Retrieved from https://open.library.ubc.ca/collections/ubctheses/831/items/1.0101387

Tate. (n.d.). Public art. https://www.tate.org.uk/art/art-terms/p/public-art

Vancouver City Council. (2019). (rep.). Making Space for Arts and Culture: Vancouver Cultural Infrastructure Plan (pp. 1–40). Vancouver, BC.