Weekly Update

Tom: This week I researched how to get 1% of tweets daily that only have geo coordinates so we can properly use the united states heat map. I need to see how we accessed the tweet stream and change the statuses/filter setting to look for only those tweets with geo coordinates.

Logan:

Kelly: Finished writing the script to get frequency of hashtags within the 1% available, currently working on the preprocessing of the text to get just the "plain text" to analyze. the way that the tweets and their entities are given to us in is lists of lists of lists. to get just the text we need to compare the all the lists and pull out the hashtags, urls, user mentions, "RT" and emojiis. the next step is to analyze the lists of plain text to get sentiment score using naive bayes classifer built into NLTK.

What to Do Next

with most of the data processing done, the next step for us is to visualize what we have effectively. as a team, we are planning on going to the CS department Colloquium on Monday.

• AttitudeAnalytiks