Differences

This shows you the differences between two versions of the page.

Link to this comparison view

05_13_2019 [2019/05/19 18:07] (current)
cynthia.kahn created
Line 1: Line 1:
 +===== Week Ending May 19 =====
  
 +==== Hurricane Analysis Using Tweets Geolocation ====
 +
 +** Completed Tasks **
 +
 +  -    plot all the counties evacuated (mandatory one color, voluntary another color) for Hurricane Irma in Georgia/​Florida
 +  -    plot all the counties evacuated (mandatory one color, voluntary another color) for Hurricane Maria in Puerto Rico.
 +  -    For all the tweets which are geolocated on the tweets database, we need to resolve their geocoordinates to counties. ​
 +  -    Find all people that tweeted one full month he hurricane start date and tweeted more than once from any given county. This will give us a list of people that '​live'​ in the hurricane area.
 +  -    Find all people that lived in the area (see before) that tweeted during the hurricane dates (here the filter is not the county, but the dates and the people found in the previous step).
 +  -    Find all tweets from people that lived in the area, which tweeted outside (and inside) of the hurricane area. For 1, 2 and 4 weeks after the hurricane time. 
 +
 +//Code stored here// [[https://​github.com/​thepanacealab/​Hurricane-Analysis]]
 +
 +** Next Steps **
 +  - Chase all '​residents'​ and get all their tweets from the hurricane period that we don't already have.
 +  * Step 1: Using twitteR and Twitter API, find all tweets within the date range of interest from the residents.
 +     //ex: searchTwitter('​charlie sheen',​ since='​2011-03-01',​ until='​2011-03-02'​)//​ or //tweets <- userTimeline("​realDonaldTrump",​ n=200)//
 +  * Step 2: Find geolocations of each tweet in the list above
 +    //ex: [[https://​gist.github.com/​dsparks/​4329876]] //
 +  * Step 3: Convert the lat/long s to county fips and plot!
 +  - Try the location imputation algorithms to enhance the dataset - this is one of the novel things to try
 +  - Use all tweets during hurricane from within the evacuation area and do sentiment analysis to get a plot of '​sentiment'​ over day for the hurricane times (pre, during, post). ​
 +      * [[http://​dataaspirant.com/​2018/​03/​22/​twitter-sentiment-analysis-using-r/​]]
  • 05_13_2019.txt
  • Last modified: 2019/05/19 18:07
  • by cynthia.kahn