Introduction

For the final project, I decided to continue working on the theme of the Lab report2, namely the distribution of New York galleries in recent years. Since in the previous report I only referred to different databases and thus visualized the basic information, I did not achieve the goal of interactivity and including more information for comparison. Also, in the previous project2, I focused more on analyzing the reasons behind the growing trend of New York galleries, while in the final report I wanted to focus on the distribution of galleries on the map and how users interact with them. The goal of the project was to enable users to filter the target galleries according to their needs through different customization modules.

Tool

Tableau Public – a tool to simply convert spatial files into visual products and export interactive visual links.
Photoshop – a tool for making gifs and detailed visualization.
Premiere pro – a tool for cutting and editing videos.

Dataset

I still used the previous database in lab report2 for visualization, which contains the gallery name, basic information, contact information, postal code, latitude, longitude, etc.

Process 1

First, I imported the Spatial file into Tableau Public to get the distribution of galleries as points on the map. In my initial exploration, the custom modules I came up with included Buffer parameter and name search box, and tried to include interactive links to the gallery phone’s tooltip and official URL in the functionality section.

The Buffer Parameter was inspired by a tutorial webpage. I think distance is an important factor when analyzing such geographical data. The buffer parameter makes it easy to calculate the distance between two galleries, or how many neighboring galleries are within a certain range of the currently selected gallery. Such a custom module allows the user to better plan the route to visit a gallery and to visit more neighboring galleries in the same amount of time.

After selecting Create Parameter in the data field on the left, fill in the information (Fig1) and click OK to create a new parameter, then click Create field and enter the code as shown (Fig2) to implement the Buffer parameter. When you pull the Filter to 100feet, you can see the gallery distribution in the uptown map as shown in Fig3. (Fig3). In order to avoid the situation that some users do not understand the specialized vocabulary of the Buffer parameter, I edited the title in the Filter area and added an additional explanation. (Fig4)

The name search box is also one of the filter modules that I think the interactive map should have. Users can search the name of the gallery to locate the gallery on the interactive map, and this feature is friendly to users who only know the name of the gallery but don’t know other information. Therefore, I added this feature to Filter and used Widecard Match in the list form, which is shown in Fig6.
In addition to customizing the filter, I also wanted my visual model to introduce basic information about the gallery. Therefore, I came up with the Tooltip feature. Due to a large number of galleries (900+ galleries in New York City), Hover would be the more aesthetically pleasing option, so that information about a gallery would only be displayed when the mouse is over a particular dot. The first thing that came to mind was information about the gallery’s most recent paintings, which is probably the most interesting content for users. But unfortunately, tableau public does not allow for the insertion of photos in the tooltip, and I had to abandon this idea. Instead, I put the gallery name, contact information, and official website in the tooltip. (Fig7)

My initial visualization process is listed above, and I test my users for the first time.

User test 1

I conducted user research mainly with two people: Rita, a “gallery lover” who visits galleries regularly, and Steven, a curator who is active among various galleries.

After the prototype of the interactive link was completed, I first showed it to Rita and gave a basic background. After exploring it on her own, she concluded that the interaction was simple and easy to understand at this stage, meaning that no additional language or text needed to be added to assist. However, she also made some suggestions: 1. she thought that normal users would not need the phone number of a gallery when looking for gallery-related information, as they would not call directly to ask for information. 2. she also suggested whether it would be possible to filter galleries based on some other filters, as sometimes as a tourist one might not know the specific names of local galleries.

Steven, the curator, gave the following feedback: 1. Although the use of the pink Parameter and the blue gallery dots provided some visual contrast, the large number of galleries did not look good on such a light background, and he suggested adjusting the background slightly. 2. The link to my current visualization was missing this information. Finally, he asked me, “Is your goal to be interactive or to convey information?” This is a good question, I must admit, because the main goal of a visualization project must be defined and unique, and other features can be secondary, but they must not take over.

Process 2

After getting feedback from users, I decided to modify the existing prototype.

First, I redefined my project goals and decided to make it the main feature for users to be able to filter the target galleries based on a custom filter, with basic information about the galleries as a secondary feature.

Secondly, I decided to classify galleries by zip code, which is more in line with users’ search habits and can more accurately classify such a large number of galleries into smaller individuals. I added the option of zip code in Filter, and also used the list form of Widecard match, because I observed that large websites such as CVS, Zara, Target, etc. have the function of searching the surrounding stores according to the entered zip code, I decided to adapt this query to my experimental report, and the search result is shown in the figure below. (Fig8)

I then considered the information hierarchy in the background and found that the zip code was also available in the pre-defined background of the tableau public, so I decided to depict the outer contours of the areas in the background based on zip code. I also labeled the map with the codes for each region in turn in the form of “annotate area”, which made the map more readable. （Fig9)

In terms of filtering basic information about the gallery, I removed the phone numbers that potential users found unnecessary and kept only the gallery name and URL. For the presentation of the URL, I wanted to implement the function of jumping to the URL link by clicking on it. So I referred to the tutorial. (Fig10)

But unfortunately, I don’t know if it was due to version limitations or other technical reasons, this feature was never implemented, but I still simulated its operation and created a Gif in Photoshop to show it. (Gif1)

User test 2

After the improvements were made, I conducted another round of user testing. I invited two users to continue the testing, and both of them were more than satisfied with the improved visualization report, but also made some new suggestions: 1. the gallery points on the map were too monochromatic and when scaled down to a certain size could easily confuse users and make it difficult to identify the modules they wanted to explore further. 2. users who tapped into the visualization report without knowing any context could be surprised by the number of gallery points and thus lose the desire to explore further.

Process3

After getting further feedback, I modified the visualization report again. First, I assigned the gallery points with colors based on zip code to better differentiate the galleries in different regions and also provide a better visual experience for users. Secondly, I also modified the title of the worksheet to better highlight the main goal of the project. Finally, since there are over 900 points on the map, some galleries may be covered by others, so I checked the distribution again and make sure I activated the function “prevent overlap”.

Besides my users’ feedback, I also got feedback from my classmates and professor. One of my classmates suggested that I could replace the “buffer parameter” with a term more easily understood by the user, since the user may be unfamiliar with the professional term. So does my professor, who also advised that I could use a more familiar unit of distance measurement instead of “feet”. Based on the above feedback, I changed the name “buffer parameter” to “adjacent distance”, and also calculated the distance with “block” instead of feet.

If we take a closer look at the distribution of all the gallery locations in New York City, it is easy to see that the majority of galleries are concentrated in Manhattan, while the number of galleries in other areas is relatively small and sparsely distributed. The galleries in Manhattan are often several galleries gathered in a certain small area, such as the junction of 10012, 10013, (Fig11)and Century Park’s east side(Fig12), and the junction of 10001, 10011, etc. (Fig13) The reason for this is those small galleries often need to cooperate and support each other, and they often circulate paintings between galleries, so they are located in close proximity to each other to facilitate communication. Such a distribution is also conducive to the development and promotion of art parks and contributes to the art industry in New York.

Deficiencies and future improvements

Of course, there are still some shortcomings in this lab report: 1. if there are other tools that are more suitable for inserting images in the future, I hope to be able to present thumbnails of recent exhibitions/paintings on display in Hover’s Tooltip. 2. due to some difficult technical problems, the action of “go to URL” is not really implemented in the interactive links. If there are other tools that can implement this feature in the future, I will actively try them.

Overall, this project successfully achieved the goal of enabling users to filter the target galleries and get basic information in the form of a custom filter.