When using the datasets, please cite the following paper:
S. Martini-Ipi, E. Moibob, M. Perc. Link Prediction on Twitter, Submitted to PLOS ONE, 2017. 

The resources are published under the CC BY-NC-SA 4.0 licence.
 

The data is given in 4 csv files.
The delimiter is ',' and column values are quoted with '"' (note that possible instances of '"' in tweets are also quoted with '"').

The csv columns represent:
 - ordinal number of tweet in csv file
 - SHA256 hash of the user's Twitter handle
 - the time ("YEAR-MONTH-DAY HOURS:MIN:SECONDS" where HOURS are expressed in a 24h format) the tweet was created
 - the tweet where newlines are replaced with spaces.

Example of reading one of the csv files in Python 3:
 
import csv

# print only times of tweets
with open('happy OR joy OR hope etc - 9991 tweets.csv', 'r', encoding='utf-8') as csvfile:
    tweets_reader = csv.reader(csvfile, delimiter=',', quotechar='"')
    for row in tweets_reader:
        print(row[2])