voting with watson
DESCRIPTION
Concept for an iPad application to help people make informed voting decisions using IBM’s Watson. This was in response to the ask: “create an app or service that leverages Watson’s computing capabilities and services in the cloud." This project was done in collaboration with IBM Design. My teammates for this project were Emma Fagergren and Tony Pattin. This is one of the projects for Marty Siegel's Rapid Design for Slow Change course at Indiana University Bloomington.TRANSCRIPT
VOTING WITH WATSON
FOR IBM DESIGN Emma Fagergren | Tony Pattin | Angélica Rosenzweig
”Democracy cannot succeed unless those who express their choice are prepared to choose wisely. The real safeguard of democracy, therefore, is education. ” Benjamin Franklin
INTRO
DEMOCRACY
OVERVIEW
TABLE OF CONTENTS
PROBLEM 4 RESEARCH INSIGHTS 5 TARGET USERS 6 CORE 7 EXPERIENCE MAP 9 SCENARIO & WIREFRAMES 11 SYSTEM MAP 23 DATA SOURCE 24 POTENTIAL ISSUES 25 FINANCIAL 26 THANK YOU 27 APPENDIX 28
Parsing information to stay politically informed is time consuming as well as energy consuming, and this in turn can can discourage people from voting.
PROBLEM
PROBLEM
CONFIDENCE MATTERS Voting is considered “the right thing” to do, but a low confidence in understanding of political issues and parties discourages voting (1).
PRESENTATION MATTERS It is not lack of information that is the problem, it is lack of information presented in a way that people can relate to.
RESEARCH INSIGHTS
TARGET USERS Individuals who want to vote but feel like they don't know enough to make an informed decision.
TARGET USER
CORE Make complex historical and current political data transparent and digestible to aid the voting decision process.
CORE
DO THIS BY: Letting users ask Watson questions (in natural language) Synthesizing information relating to issues that the user has identified as relevant Showing how candidate’s values align with the user’s
WE ARE NOT: Telling the user who to vote for. The app should be politically neutral.
CORE
EXPERIENCE MAP
OUR FOCUS: EXPERIENCE MAP
It is the fall of 2016 and with elections coming up millions of Americans are getting ready to vote. Some are certain who they are voting for, others are still trying to decide.
SCENARIO
Marie, a recent college graduate feels that she should vote - she has grown up hearing that “it is the right thing to do”. But who should she vote for? She’s not sure and she feels that she doesn’t know enough about politics to make a good decision. She doesn’t want to vote based on the candidates advertisements, looks or gender, but on how their values align with hers.
SCENARIO
Marie follows the news and has a general idea about the candidates, but she wants to know where they stand in question that are important to her - education and the environment. Her Google searches for candidates return thousands of hits, and none really answer her questions. She doesn’t have the time it takes to make sense of the (sometimes contradicting) information, and some candidates seem to say the same thing. Frustrated, Marie leaves her computer.
SCENARIO
SCENARIO
A couple of weeks later Marie is talking to her friend Alex about the election. Alex has an interest in politics and after Marie has voiced her frustration, he mentions a new app he has found. He suggest Marie use it to find out answers about things she cares about in the upcoming election.
After listening to Alex, Marie decides to give it a go. They try out the app together and Alex helps Marie create an account. She gets to choose which political issues she is interested in. Marie also fills out where she lives.
WIREFRAMES
When her profile is created, Marie is presented with a home screen that serves as a one-stop shop for her upcoming election research. She notices the information shown relates to the issues she thought were most interesting. On the home page, Marie has access to various information: Trending Topics Media Candidates Personal Profile Watson Favorites
WIREFRAMES
Marie is interested in knowing what others have been saying about the similar topics, so she decides to click on Trending. Here, Marie can see questions that other people have asked “Watson”. She browses through the questions that other people have asked. Some seem irrelevant or strange to her, but other questions peak her interest and she clicks on “See More” to read the answers generated by Watson.
WIREFRAMES
Marie decides to ask Watson a question herself. She wants to know which candidate in her state is most passionate about creating better schools, and starts typing in her question.
WIREFRAMES
Watson then presents Marie with results from the search. The top answer is presented with a high confidence level; the other two results are presented with a medium confidence level. She saves this question, then continues to explore other questions of her interest.
WIREFRAMES
Later that week Marie and Alex are watching a TV-debate between two senators. Marie wants to know more about the senators and opens up the app. After giving her opinion on a number of questions, Marie is able tosee to which degree the candidates values align with her own. She likes this summary; it gives her information beyond what she has learned from the debate.
SCENARIO
Marie explores the “Candidates Page”. With a 90% match, she has found out that the one of the senators aligns with most of her values and interest.
SCENARIO
A couple of weeks passes and soon it is time to vote. Marie feels that this time she has tried her best to make an informed decision, and for once she will stay up to follow the results on TV.* *Who Marie ended up voting for is confidential.
SCENARIO
SYSTEM MAP
DATA SOURCE
CORPUS DATA Historical political records News feeds Current legislation Political parties site Wikipedia Political Blogs
DIFFERENT CANDIDATES Watson is likely to find more information for experienced candidates than new candidates. This might affect confidence levels.
CONFIDENCE LEVELS Use Watson’s Confidence levels to avoid deception. The nature of political information is often biased, but by looking at many different sources, Watson triangulate between different sources to create meaning and show confidence levels in the response
POTENTIAL ISSUES
SELLING DATA After a period of use, the app will have collected user data that will be of interest to politicians and campaign managers. Selling this data would generate revenue.
POTENTIAL SPONSORS Independent civic organizations Project Vote Smart Federal Election Commission (FEC)
FINANCIAL
THANK YOU
SPECIAL THANKS TO:
IBM Design Tarun Gangwani
Chung-Ching Huang Brandon Le
Mark Marrara Marty Siegel
HCI/d 2014 Cohort
APPENDIX
RESEARCH (1) Youth Federal Election voting intention http://www.whitlam.org/__data/assets/pdf_file/0005/507803/Newspoll_Analysis_Update_Election_Poll_Update_Aug_2013.pdf @ 20 Nov. 2013 (2) http://photos.state.gov/libraries/amgov/30145/publications-english/USA_Elections_InBrief.pdf @ 20 Nov. 2013
INSPIRATION FROM EXEMPLARS
APPENDIX
Politico (app) One-stop shop for coverage of the president, Congress, 2014 Midterm elections, and 2016 presidential race
Settle It! (app) Fact checking site for political information
OUR PROCESS Understanding Watson Brainstorming problem spaces Research: Understanding the voting process Personas Sketching and problem exploration User case scenarios High-fidelity wireframes and graphics Crafting deliverables
APPENDIX