Class Data Visualizations

Sentiment analysis of primary sources about Chinese cuisine in the United States. 

Sentiment Analysis

Process

I compiled all of our group members different content sets on Digital Scholar Lab into one, containing 57 documents. I used the Sentiment Analysis tool on DSL to analyze our data. I ran two analyses, one without using the cleaning configuration and another with using the cleaning configuration. It seemed to me that there was not much difference between the two results.

Analysis and Consideration

The visualization of the sentiment analysis for our sources confirmed some issues, such as the negativity surrounding Chinese food resulting from the Chinese restaurant syndrome in the late 1960s. However, we were not too convinced about the results and questioned how this tool gave sentiment scores to documents in our contentset.  since some of the documents, dating back to the late 19th century were clearly more negative in nature than some of later documents surrounding the Chinese restaurant syndrome, yet they received a more positive score in the analysis. Overall, I think our timescope was too large and the documents spanning that timescope were too few to be truly representative of anything conclusive.

Chinese Restaurant Syndrome Word Cloud

Chinese Restaurant Syndrome Word Cloud

Process

I downloaded thirteen documents that were in our "Chinese Restaurant Syndrome" content set on Digital Scholar Lab and uploaded them all onto Voyant as a single txt file. Then Voyant generated the word cloud. 

Analysis & Considerations

The most frequent terms were "Chinese", "msg", "food, "restaurant", "syndrome", and "glutamate". Which was not surprising since most of our documents were centered around the effect that MSG had on Chinese food and how both of them together created the Chinese Restaurant Syndrome. Some other words worth noting are: "health", "disease", "pain", "headache", and "cancer". These words are interesting because they are all health-related terms and most of them have negative connotations. One looking at this word cloud without knowing what Chinese Restaurant Syndrome is can conclude that it is a negative thing. 

Geoparsing Articles on Nixon's 1972 Visit to China

Screen Capture of Geoparsing Visualization

View on Google Fusion Tables →

Process

I downloaded six documents from Digital Scholar Lab, as well as four documents found from other sources, and concatenated the plain text into a single .txt document. All documents are about President Nixon's diplomatic visit to China in 1972, a trip which led to a spike in Chinese food popularity in the United States. I used Clavin Geoparser to extract location data from the documents, then I uploaded the data to Google Fusion Tables, which mapped the coordinates.

Analysis & Considerations

I found it really interesting that location data was concentrated in other locations besides China and the United States, and included seemingly unrelated locations, including Pakistan, Guam, Paris, Turkey, and Cambodia. This leads me to question the accuracy of Clavin Geoparsing -- but I also wonder in what context these locations were mentioned. Unfortunately, it is difficult for me to check the tool's accuracy, because Clavin provides no data besides the location name and its coordinates. If it showed the frequency of location occurrences or where in the text the location was found, I would be able to confirm these findings, and learn why the locations may have been mentioned.

It is also worth noting that I only ran ten documents through Clavin, so the results may not be entirely representative of the subject area.

Class Data Visualizations