Description:These scripts are part of a larger project focusing on examining political elites in Colombia on Twitter. The code found in this file can be used to gather locations of a Twitter user's 'friends' (i.e. who they themselves follow) and map all of the 'friends' that have put some sort of location into their profile info.

IDEs used:
Python: Jupyter Notebook, Python 3.6.1
R: RStudio, R Version 3.4.3 Kite-Eating-Tree
Packages used are:
Python: tweepy, json, csv, pandas, datetime, time
R: dplyr, ggmap, ggplot2, maps, stringr

Order of running files is numerical.
1_twitter_api.ipynb
2_data_cleaning.R
3_visualization.R
4_saving_results.R

Go to developer.twitter.com to register an app and receive the 4 needed keys to access the API.
Create an empty (truly empty!) csv file in sublime or nano to work with for this script.
Run script! Refer to inline comments for explanations and logic.
Expected processing time is around 3 hours

Read in the dataset produced from the previous script.
Clean and manipulate the data (see inline comments for explanations.)
Begin 3rd script when the main dictionary and all locations lists have been created. Dictionary uses Google Maps geocoding API and requires another app and keys.
Note: When calling the geocode api, the rate limit is 2500 requests per day. Make sure you're calling these data right the first time Expected processing time is around 30 minutes

Using ggplot2, ggmap, and maps, create map charts (currently: 10)
Expected processing time is around 10 minutes

In the loop, ggsave each chart as pdf. (Best to import in laTEX)
Expected processing time is around 10 minutes