Text Mining in Jeb Bush’s Email and Social Network
Download as pptx, pdf2 likes764 views
The document analyzes Jeb Bush's email and social media data from 1999 to 2006 to identify his key interests and voter sentiments. It highlights significant patterns in email communication around education and environmental issues while revealing negative sentiments towards him in social media discussions. The analysis utilized Python, Weka, and SPSS for data processing, visualization, and sentiment analysis.
Text Mining in Jeb Bush’s Email and Social Network
- 1. Gap Analysis in Jeb Bush’s Email and Social Network Team Member: Yi Chun Chien (Nancy), Jing Fan, Pei yun Yeh, Tianmiao Zhou 1 Class : Text Mining Professor : Yilu Zhou
- 2. Jeb Bush Jeb Bush Younger Brother George Bush Sr. Father 43rd U.S. president George Bush Older Brother 41st U.S. president ? Governor of Florida (1999-2007) ? Member of the Bush political families ? Converted to Catholicism (Under his wife’s influence) ? Speaks Spanish ? Strongly supports nation's immigration laws ? Close mentor to Marco Rubio (Also running for President 2016) 2
- 3. ?Open data ?Extract context ?Summarize Statistics Data Collection ?Customized stop words ?Word Matrix ?Delete redundancy feature manually Preprocessing ?Visualization ?K-Mean Clustering Analysis ? Purpose: (a) Find out which topic or field Jeb Bush cares the most from his email contents in 1999-2006 (b) compare the voters’ opinions from Twitter ? Why Jeb Bush released Email: “In the spirit of transparency, I am posting the emails of my governorship here” ? Total ~2.7 million mails (Average 949 a day!) ? Tools: Python and Weka Email and Twitter Analysis Email Analysis 3
- 4. Email Count vs. Quarter over Years Presidential Election 1999 2000 2001 2002 2003 2004 2005 2006 EmailCount Quarter Presidential Election 4 ? 1st and 2nd presidential election has different email amount EmailCount Quarter Quarter Quarter Quarter Quarter Quarter Quarter Florida educational Performance issue
- 5. 2nd Run Period 2nd Run Period 1st run Period Email Count: Weekday vs. Weekend ? (a) Less mails on weekend (b) 2004 has different pattern than other years ? Compared to his 1st and 2nd run, the email amount increases from 409/day -> 1,283/day 1999 2000 2001 2002 2003 2004 2005 2006 EmailCount Sunday ~ Saturday 5 Sunday ~ Saturday Sunday ~ Saturday Sunday ~ Saturday Sunday ~ Saturday Sunday ~ Saturday Sunday ~ Saturday Sunday ~ Saturday EmailCount
- 6. Word Clustering Result 1999 ? Children - Florida ? Schools - Governor ? Education - Bill ? Parents - Life 2000 ? Gas - Children ? Oil - Health ? Draft - Law ? OCS - Water 2001 ? Gas - Water ? Oil - Bill ? Leasing - School ? Ocs - Health 2002 ? Everglades - Children ? Monroe - Education ? Protection - Family ? Endangered - Manatees 2003 ? Handicapped - Custody ? Exhausted - Constitution ? Dismantled - Protect ? Murder - Judge 2004 ? Sanctuary - Fight ? Manatee Protection - Brother ? Legislation - Respiratory ? Inhabitants - Church 2005 ? Education - Kathleen ? Manatee - Secretary ? Miami ? Budget 2006 ? Manatee - School ? Protection - Children ? Waters ? Species 6 Cluster 1 Cluster 2 Cluster 1 Cluster 2 ? Use K-mean clustering to group all key words into 2 clusters ? Education and Environment are the 2 hot topics
- 7. Social Network Mining ? 1779 Tweets about Jeb Bush from Twitter ? Excel Format Data Collection ? SPSS Text Analytics Tool ? Customized library ? Three Types: Positive, Negative and Neutral Preprocessing ? Key Word Analysis ? Association Analysis Sentiment Analysis 7
- 8. Sentiment Analysis Result 8 ? When people mention “president” in #JebBush, the word also associate strongly with “brother” and “thank”
- 9. Sentiment Analysis Result (Conti.) 9 ? When people mention “gop” (Republican National Committee) in #JebBush, the word is related greatly to “immigrants”
- 10. Sentiment Analysis Result (Conti.) 10 ? When people mention “jebbush” in #JebBush, the word also associates with many negative words
- 11. Social Media Result ? Two typical hash tag: #NoMoreBush #StopHillary ? Tweets show people do not like Jeb Bush and neither Hillary Clinton. ? When people talked about Jeb Bush, they mentioned George Bush at the same time. ? People prefer to criticize politicians than praise them 11
- 12. Conclusion Year1999 2000 2001 2002 2003 2004 2005 Education Governor Education Environment Education Environment Education Environment Healthcare Environment President Education Environment Scandal Education Environment 2006 From Analysis ? Most of his email content match his later policy ? Has a powerful background but long way on his Presidential Election Future Work ? Lexical analysis for Email (EX: pre-define keywords) ? Build larger library (In SPSS) 12 Topic
Editor's Notes
- #4: Methodology: String to word vector: word frequency Define Stop words Manually delete meaningless words (EX: verb, adv; 1000 features -> ~200 features) Use K-means clustering to group these features
- #5: December 11, 2000. He was the Florida governor and the Supreme Court was hearing a historic Florida case. It happened to be called?Bush v. Gore, and it would determine whether his older brother, George, would win his state’s electoral votes and become the next president Read more:?http://www.politico.com/magazine/story/2015/03/jeb-bush-everglades-115655.html#ixzz3XufCMKWP
- #7: http://www.ontheissues.org/Jeb_Bush.htm =====================================?Environment?===================================== Support State Revolving Loan Fund for flexible Clean Water. (Aug 2001) Drilling in Gulf of Mexico hurts Florida tourism industry. (Jan 2001) New marine sanctuary to protect the Florida Keys. (Mar 2004) =====================================?Families & Children?=====================================? No mandated child safety seats. (Jun 2001) Parental consent over government intrusions into families. (May 2001) Encourage fathers' participation in child-raising. (Sep 2001) Federal funds & state involvement in fatherhood initiatives. (Aug 2001) Increase KidCare; increase developmentally disabled services. (Jan 2002) No Place Like Home initiative: find families for DCF kids. (Mar 2004) =====================================?Free Trade?=====================================? Advocated Miami as HQ for Free Trade Area of the Americas. (Nov 2003) =====================================?Healthcare?=====================================? 2003: http://www.inclusiondaily.com/archives/04/09/23.htm Florida Supreme Court Rules "Terri's Law" Is Unconstitutional; () Disability Advocates Disappointed But Resolved To Continue Legal Battle =====================================?2005?=====================================? Katherine Harris?(born April 5, 1957) is a former?Secretary of State of Florida?and former member of the?United States House of Representatives. A?Republican, Harris won the?2002 election?to represent?Florida's 13th congressional district?in the U.S. House of Representatives. She held that post from 2003 to 2007. Harris lost the November 7, 2006, election to represent Florida in the?United States Senate. Florida Senate and Riscorp[edit] Harris played a prominent role in introducing the CEO of Riscorp, William Griffin, (with whom she had a close personal relationship) to various Florida legislators. In the 1994 state senate election,?Sarasota-based Riscorp, Inc. made illegal contributions totaling $400,000 to dozens of political candidates and committees,[11]?including $20,600 to the Harris campaign.[12] Two years later, in 1996, Harris sponsored a bill "to block Riscorp competitors from getting a greater share of Florida?workers' compensation?market, [and] also pushed a proposal that would hurt a particular competitor."[11]?This issue later emerged during her campaign for Florida Secretary of State in 1998. William Griffin, eventually pled guilty to illegal campaign donations among allegations of other serious wrongdoing at Riscorp and served prison time in 1998. The election of?Jeb Bush?as governor of Florida was a major factor in stopping further investigation into the Riscorp scandal. According to a SunHerald column from June, 2005, "Harris denied any knowledge of the scheme, was never charged with any crime and was cleared of wrongdoing by a state investigator."[13]?This view was criticized strongly by other investigators involved in the Riscorp prosecution.
- #8: How many key words? What’s Similarity?
- #13: During his eight years as governor, Bush was credited with initiating improvements in the environment, as well as reforming the education system.[3][4]