AwardPuzzle

Year: 2014–16
Role: Research · Platform · Publication
Team: Fan Xiang · Shunshan Zhu
Source: National Exhibition of Fine Arts catalogues · Tsinghua University Library · National Library of China
Includes: Interactive platform · Video · Paper · Talk

AwardPuzzle is an interactive visualization platform for exploring commonalities, trends, and differences among 2,276 prize-winning oil paintings from seven editions of China’s National Exhibition of Fine Arts. It lets the public compare thirty years of works without reducing the paintings to abstract marks.

Original project record

View archived page ↗

Interactive public platform released in May 2016 after the first research report circulated in January 2015.
Data collection: Wenshi Luo · Xinyan Huang · Bo Jin · Manhua Zhang · Pingyun Liu. Image preprocessing: Shulun Ye · Haoji Lin · Yu Zhang.
Longlisted for the Kantar Information is Beautiful Awards 2016 in Interactive Visualization.

Project extension

Image similarity analysis

An extension of AwardPuzzle, this interactive visualization connects paintings from the National Art Exhibition through image similarity. Explore 349 images and 443 connections to compare visual relationships across the works.

Open the image similarity analysis websitehttps://xiangfan.ink/zeelab/AP/APsim.html

Opens in a new tab · Drag nodes and zoom to explore

From dispute to evidence

The National Exhibition of Fine Arts is presented as a survey of the country’s highest artistic achievement, yet critics have questioned recurring subjects, colors, and compositions. AwardPuzzle asks how these impressions can be examined when no individual can remember and compare thousands of paintings across thirty years.

Building a comparable image archive

The organizers did not provide a complete online record. The team collected and scanned catalogues covering seven exhibitions from 1984 to 2014, then structured each work by artist, title, region, year, dimensions, color values, and manually assigned visual tags.

Keeping the paintings visible

Instead of replacing paintings with bars or dots, the interface uses each image as its own data mark. This direct visualization keeps the research object visible while allowing more than two thousand images to be observed as one field. Hover, zoom, tagging, and filtering connect distant and close reading, making observations available for inspection rather than offering a single verdict.

Five ways to compare

A linear timeline shows changes in volume; a polar color field maps median hue and brightness; a width–height plot compares canvas dimensions; a geographic view groups artists by region; and an artist view gathers repeat winners. Animated transitions let the paintings retain their identity as the organizing model changes.

Questions generate new views

Development moved through a repeated cycle of question, visualization, and new question. Similar paintings with the same title prompted an artist view; debates about increasingly large canvases prompted a dimensions view; transitions between views revealed further relationships among size, region, and time.

Patterns made inspectable

The color view reveals a persistent concentration of bright yellow and red tones. The size view shows canvases expanding after 1999 and reaching a high point in the eleventh exhibition, followed by a uniform upper limit in 2014. Gathering works by artist also makes repeated themes and closely related compositions available for direct comparison.

A public research platform

An initial report circulated through WeChat and Weibo before the interactive platform opened in 2016. Public use demonstrated both the value of independent exploration and a practical limitation: loading thousands of images could take too long. The project treats interface performance, user attention, and interpretation as part of visual research.

What the map cannot claim

The archive represents works selected by one national exhibition system, not Chinese oil painting as a whole. Median color simplifies the material complexity of paint, and printed catalogues introduce color variation through paper, ink, preservation, and scanning. The visual patterns therefore describe this collection and its selection process—not a universal history of art.