Project Overview
By utilizing text mining tools, insights can be gained around the sentiment that internet users are communicating with each other regarding specific topics. Using Brand24.com’s online text mining tools, an analysis was performed on different intermountain west cities to determine the general sentiment, either positive or negative, that people were expressing when discussing those cities. Spokane, Salt Lake City, and Boise are all intermountain west cities with similar population sizes, so they represented a good sample of cities to compare. From May 1st to May 31st, different online sources were mined for text including sources such as news articles, website posts, blogs, forum posts, podcasts, and videos. Starting May 20th to May 31st, Twitter and Facebook posts were also included with Brand24’s two-week trial. Spokane and Salt Lake City were nearly identical at the top until the last day of the month when protests erupted across the country. This caused Salt Lake City to have more total mentions with 6,795, but they also had the most negative sentiment across the cities due to the more violent and destructive turn that the protests took there. Boise may have had the least mentions, but it did have the highest positive sentiment total of 73.3% positive.
Analysis Approach
Initially the sentiment analysis was only run for the City of Spokane. However, for the results to provide more insight, a comparison analysis was also run to determine how these faired against similar cities. While comparing cities can sometimes be difficult due to size, climate, and location differences for example, there are a few other cities that are nearly identical in size and climate that would be good candidates to compare against Spokane, Washington. Boise, Idaho and Salt Lake City, Utah are also cities with populations around 200,000 in the intermountain west region. The online text mining trackers were setup for those cities as well. Initial results showed Spokane to have the most mentions, as well as the highest positive mentions. These initial results though did not include social media, which can be a large factor in actual sentiment. The additional two-week social media trial was then activated to track Twitter and Facebook posts as well. This increased the total documents being analyzed by over 25%. It also highlighted the spike in the final day of the month that upon inspection were due to the protests that turned violent in Salt Lake City.
Takeaways
The initial hypothesis was that Salt Lake City would have the highest mentions and Boise would have the lowest. While this was the case, it was surprising that if excluding the last day of the month where an extreme spike occurred that was generally negative for Salt Lake City, Spokane was nearly identical to total mentions and sentiment than Salt Lake City. Spokane has a social media tag #spokanedoesntsuck that indicates a general perception that the city has a negative sentiment, however, this analysis show that the general sentiment is positive and even more positive than the more well-known Salt Lake City. Boise may have had the least amount of mentions online, however, they had the most positive sentiment of any of the cities. Since the cities are close in size, and all are outdoor focused, it may be interesting to investigate the impact of increased social media use in the other cities leading to more dissatisfied citizens.
Many of the social media documents came from Twitter and Reddit. This indicates that these may be better outlets for the local governments to analyze further for more citizen feedback. Overall, for each of these cities, the overall sentiment was positive. However, as was seen in Salt Lake City, it only takes a single event for sentiment to change quickly. It would be useful for cities to perform text mining and social media monitoring activities in a more robust way to ensure they are following sentiment trends and discovering the true passions of the community. Often the squeaky wheel gets the grease, however, with detailed analysis of the community sentiment, effort can be applied to things that improve life for citizens in a more transparent and communicative way.
Author: Logan Callen
References
Marta. 2020. “A comprehensive guide to social media analysis.” Brand24, Accessed May 24, 2020. https://brand24.com/blog/guide-to-social-media-analysis/
Provost, Foster and Tom Fawcett. 2013. Data Science for Business. 2nd Edition. California: O’Reilly Media, Inc.
Wikipedia. “Boise, Idaho.” Accessed May 24, 2020. https://en.wikipedia.org/wiki/Spokane%2C_Washington
Wikipedia. “Salt Lake City” Accessed May 24, 2020. https://en.wikipedia.org/wiki/Spokane%2C_Washington
Wikipedia. “Spokane, Washington.” Accessed May 24, 2020. https://en.wikipedia.org/wiki/Spokane%2C_Washington
0 comments on “Text Mining Sentiment Analysis – Intermountain West Cities”Add yours →