Our visualization tracks the sentiment of some newspapers towards the Presidential candidates from the time-period starting just after the first debate and just before the election night.

Morgan - Hacker and developer - Python scripting to scrape newspaper sites and interact with NYTimes and NPR APIs to get relevant article URLs formatted in JSON for Sentiment Analysis API calls. Manually made JSON data for all SF Gate articles since there is no API available Tried unsuccessfully to use Google Custom Search API to scrape various newspapers

Bharathkumar Gunasekaran - Hacker & Developer - Researched various sentiment analysis tools; implemented sentiment analysis for all articles with "Alchemy API". Formatted data to be CSV for use in D3 visualization

Chan Kim - Hacker & Designer - Created javascript visualization that graphs Presidential candidate sentiment in various newspapers in the weeks before the election with special reference to the debates. Also created the HTML/CSS front-end of the site.

Python
HTML/CSS
Javascript
D3
Alchemy API
NT TImes API
NPR API
Google Search API

Live URL - http://people.ischool.berkeley.edu/~chan/iolab/sentiment_times/ GIT Repo : https://github.com/morganwallace1/IOLab-Project-3

We tested on Chrome and Firefox

1. When there are multiple articles on a particular day, we tried to average the scores and have a single score for the day. But due to time constraints, we were ont able to implement this