When the "data visualization" full of science and technology meets the classical Tang poetry and Song poetry, what kind of spark will it spark?
Recently, a group of visual data works, Song Ci, Where to Draw the World (hereinafter referred to as Song Ci) and Tang Nv Poets Group (hereinafter referred to as Tang Poetry), have been screened in the circle of friends. The work was completed in half a year by the State Key Laboratory of CAD&CG of Zhejiang University and the Data and News Department of Xinhuanet.
The team analyzed 55,000 Tang poems and 21,000 Song poems, and interpreted classical Tang poems and Song poems with big data technology, and unexpectedly found many hidden information.
Big data display
The place where Su Dongpo went the most was Hangzhou.
What does this work look like? Qianbao reporter opened the web version of Song Ci.
There are many blank spaces, ink illustrations and elegant color matching style. The style of this work is "close to landscape painting as a whole".
It is understood that "Song Ci" takes "Song Ci" as a sample. In order to complete this interpretation, the team analyzed nearly 21,000 ci poems, nearly 1,330 poets and nearly 1,300 epigrams. The basis of the works of Tang Poetry is the data analysis of 55,000 Tang poems.
The reporter observed that the web version of Song Ci is composed of Shi Kongtu linked by the poet’s trace map and the life map of the year, as well as the cloud map, image emotion map and prosodic map of Song Ci.
In the chronological life map, the reporter selected the broken line representing "Su Shi". According to the diagram, a broken line of "first leveling, then rising and leveling" was displayed, which tried to reflect the ups and downs of Su Shi’s career.
On the track map linked with the life map, brown points with different sizes appear synchronously on the map, which are connected by lines, and the size of each point is determined by the number of times Su Shi set foot. This shows the trajectory of Su Shi’s life. From the point line diagram, Su Shi’s footprints are almost all over the territory of the Song Dynasty. Among them, the biggest place is Hangzhou City, which shows that Hangzhou is the place he visits most.
The word "east wind"
It appeared 1264 times in Song Ci.
With the drop-down of the page, what comes into view is the "cloud picture of words" in Song Ci. According to the analysis of the number of words used in Song Ci, the more times they are used, the bigger the font size, the darker the color and the more centered the position. The reporter saw that the word in the middle was "Dongfeng", which was used 1264 times. Followed by "where", it was used 1157 times. The third place is "human world", which appeared 1061 times in Song Ci.
"We used to understand the poems of the Song Dynasty and the Tang Dynasty, and more of them were understood and appreciated separately. This study allows us to find the hidden information behind the poems from the big data level." Theway, design director of the State Key Laboratory of CAD&CG of Zhejiang University, told Qianbao reporter.
The research lasted for half a year, and the works of Song Ci and Tang Poetry produced by Zhejiang University team and Xinhua News Department were all presented in the form of web pages, which contained quite rich information. Among them, the most informative and complicated work is Song Ci.
"In the media industry, such mature visual data news works on traditional cultural themes are still rare." Theway said that this is also the first attempt by the visualization team of Zhejiang University.
The poet mentioned "wine" in his works.
Half is thinking, 30% is happy.
The team not only analyzed the superficial information of the text of Song Ci, but also deeply explored the image meaning expressed by Song Ci and integrated it into an image emotion map.
30 common words such as "moon" and "wine" are selected in the image emotion map, represented by 24 prolific poets such as Su Shi and Li Qingzhao. Through the analysis of big data, the emotions expressed by these image words are obtained, and the emotions are divided into five types — — "Think about emotions", and then use pie charts to show the proportion of different emotions expressed by various words.
For example, when poets write "wine", nearly half of the images they want to express are nostalgia and thinking. Lu You wrote "Red Crispy Hands, Huangteng Wine", or Yan Shu wrote "A new song and a glass of wine, the weather was old last year", all of which are reminiscing about old friends and thinking about life. There are nearly 30%, which is similar to Zhu Dunru’s "a cup of wine is full every day, and flowers bloom in the small garden".
Then, how did big data technology observe the poet’s mood at that time from the lines of Song Ci?
First of all, the team needs to sort out the typical images that basically only represent a certain emotion. Theway said that in order to be more accurate, the team specially invited Dr. Hu Qiuyan from Zhejiang University College of Literature to check.
Pan Rusheng, who is in charge of data analysis and front-end development, told reporters that they will use big data to analyze the context, calculate the probability that the word belongs to a certain emotion according to typical images, and get the emotion that the poet is most likely to express.
To put it simply, for example, the poet Zhang Zai wrote in "The Old Cypress Courtyard of Xinglong Temple": "The peony blossoms in the south and the peony blossoms in the north, and the young people look for fragrance several times. Only the old cypress tree in your family seems to have never come. " Among them, "pine and cypress" expresses an emotion of "remembering". In the context, it can be concluded that what "Peony" and "Spring Breeze" want to convey is also "thinking".
Through visual data presentation
Make Tang poetry and Song poetry truly popular and easy to understand
When asked about the difficulty of this research, Theway first mentioned the choice of charts. In order to find the most suitable data presentation, many charts are easy to draft.
Appropriate charts should not only be beautiful, but also cover the information that needs to be presented. At the same time, they should be intuitive and interact with readers smoothly, which really makes the team spend a lot of time. According to Theway, the team tried to use the view of "small mountain peak" to express the cadence of words, but considering that the overlapping of images affected the impression and was not conducive to the placement of imagery images, they finally gave up.
"People are visual, and the visual form of popular science means can make obscure ancient poems easy to understand, and let popular science get rid of preaching or boring stereotypes, thus playing a role in promoting traditional culture." Chen Wei, deputy dean of the School of Computer Science and Technology of Zhejiang University, said.
The orientation of this research is popular science, so the object of analysis is mainly the most basic content of Tang poetry and Song poetry. Theway said: "This product is not made to draw a certain conclusion, but to provide people with an interesting tool to explore Tang poetry and Song poetry." Therefore, more interesting conclusions remain to be discovered by readers.
