Description
In this assignment, we are going to use tweets as our datasets. The general idea is to
review basic concepts of Python, like lists and dictionaries, and make sure that you are
all familiarized with those data structures.
The datasets are csv files that store, in each line, a public tweet from a user:
@ al y s s a p r a g e r , Thu Sep 18 1 8: 0 3: 4 0 EDT 2 0 1 4 , −7 3. 9 8 5 3 3 , 4 0. 7 5 9 3 3 7 , #NewYork
, # Tim e s S q u a r e
The first column is the screen name of the user in Twitter (alyssaprager in this
case). The second column is the date and time of the tweet (Thu Sep 18 18:03:40 EDT
2014). The third and fourth columns are the longitude and latitude of the user when he
tweeted. Starting from the fifth column, we have the hashtags used by the user in his
tweet; in this example, the user tweeted two hashtags: #NewYork and #TimesSquare.
The number of hashtags is arbitrary per line.
All the tweets were captured using Twitter’s public streaming API. More informations here: https://dev.twitter.com/streaming/public.
All outputs need to be exactly as in the examples. Pay attention to date formats,
spaces and punctuation. In this assignment you’ll get zero if the output is different
from the expected. A tip: you can use a tool to compare your output with the sample
output provided. In unix systems you can use the diff command to check your output
against the expected one. For windows, you can do this using the FC command.
For all problems, your solution should receive command line parameters with the
input files names, so it should not depend on internet connection to run. Download
the sample files to your computer and point to them to execute your solutions.
You cannot use Pandas in this assignment.
Problem 1
Create a Python file that reads each line of the dataset http://vgc.poly.edu/
projects/gx5003-fall2014/week3/lab/data/sample_data_problem_
1
1.txt and computes two things: number of tweets in the csv file and range of dates.
Your output should be in the format:
The output of the sample is available at http://vgc.poly.edu/projects/
gx5003-fall2014/week3/lab/data/sample_output_problem_1.txt.
Note: attention to the date format, spaces and punctuation.
Problem 2
Compute the number of unique users in the dataset and also the range of dates in the
dataset. For this problem, use the dataset http://vgc.poly.edu/projects/
gx5003-fall2014/week3/lab/data/sample_data_problem_2.txt. Your
output should be in the format:
> p yt h o n p r o blem 2 . py s am pl e _ d at a _ p r o bl em _ 2 . t x t
49 u s e r s t w e et e d betwee n S e pt em b e r 19 2 0 1 4 , 2 1: 0 0: 1 8 and S e pt em b e r 19
2 0 1 4 , 2 1 : 0 0: 2 9
The output of the sample is available at http://vgc.poly.edu/projects/
gx5003-fall2014/week3/lab/data/sample_output_problem_2.txt.
Problem 3
Given two datasets, output all hashtags that happen in both datasets. The hashtags must
be unique (in other words, no two hashtags outputed can be equal), and they are sorted
in lexicographical order. For this problem, use the datasets http://vgc.poly.
edu/projects/gx5003-fall2014/week3/lab/data/sample_data_problem_
3_1.txt and http://vgc.poly.edu/projects/gx5003-fall2014/week3/
lab/data/sample_data_problem_3_2.txt. Your output should be a list of
all hashtags sorted in lexicographical order, one hashtag per line.
The output of the sample is available at http://vgc.poly.edu/projects/
gx5003-fall2014/week3/lab/data/sample_output_problem_3.txt.
Problem 4
Compute the 10 most popular hashtags in the dataset. The hashtags must be sorted by
popularity. For this problem, use the dataset http://vgc.poly.edu/projects/
gx5003-fall2014/week3/lab/data/sample_data_problem_4.txt. Your
output should be a list of all hashtags sorted in popularity, with their count value, one
hashtag per line. If the count of two hashtags are the same, then they should be sorted
in lexicographical order.
The output of the sample is available at http://vgc.poly.edu/projects/
gx5003-fall2014/week3/lab/data/sample_output_problem_4.txt.
Problem 5
Compute the date and time when most of the tweets happened. Consider seconds
as the time granularity here. For instance, if 10 tweets happened on Thu Sep 18
18:03:30 EDT 2014, and 11 tweets happened on Thu Sep 18 18:03:31 EDT 2014,
2
then Thu Sep 18 18:03:31 EDT 2014 is the date and time when most of the tweets happened. For this problem, use the dataset http://vgc.poly.edu/projects/
gx5003-fall2014/week3/lab/data/sample_data_problem_5.txt. Your
output should be in the format:
> p yt h o n p r o blem 5 . py s am pl e _ d at a _ p r o bl em _ 5 . t x t
S e pt em b e r 19 2 0 1 4 , 2 1 : 3 9: 1 9 wit h 13 t w e e t s
The output of the sample is available at http://vgc.poly.edu/projects/
gx5003-fall2014/week3/lab/data/sample_output_problem_5.txt.
Problem 6
Now, consider that the granularity of the time is hours. For instance, if 10 tweets happened on Thu Sep 18 18:03:30 EDT 2014, 11 tweets happened on Thu Sep 18 18:03:31
EDT 2014, and 15 tweets happened on Thu Sep 18 19:05:30 EDT 2014. Then, 18:00 of
Sep 18 is the time and date when most of the tweets happened. For this problem, use the
dataset http://vgc.poly.edu/projects/gx5003-fall2014/week3/lab/
data/sample_data_problem_6.txt. Your output should be in the format:
> p yt h o n p r o blem 6 . py s am pl e _ d at a _ p r o bl em _ 6 . t x t
S e pt em b e r 19 2 0 1 4 , 20 h wit h 15996 t w e e t s
Note that we are only outputting the month, day, year and hour, because our time
granularity is hours.
The output of the sample is available at http://vgc.poly.edu/projects/
gx5003-fall2014/week3/lab/data/sample_output_problem_6.txt.
Problem 7
Now, output the user who tweeted the most in the dataset and also the date and time
range of the dataset. Note that it doesn’t matter if the user tweeted 1 hashtag or
10 hashtags; in this problem, we are ignoring the number of hashtags used by the
user. For this problem, use the dataset http://vgc.poly.edu/projects/
gx5003-fall2014/week3/lab/data/sample_data_problem_7.txt. Your
output should be in the format:
> p yt h o n p r o blem 7 . py s am pl e _ d at a _ p r o bl em _ 7 . t x t
@t r e n di n ali a U S t w e et e d t h e most
D a t a s e t r a n g e : S e pt em b e r 19 2 0 1 4 , 2 1: 0 0: 1 8 and S e pt em b e r 19 2 0 1 4 ,
2 1: 0 0: 2 9
If two users have the same number of tweets, you should use lexicographical order
to break the tie.
The output of the sample is available at http://vgc.poly.edu/projects/
gx5003-fall2014/week3/lab/data/sample_output_problem_7.txt.
Problem 8
In our datasets, we also have the latitude and longitude of the tweet. We will calculate
now the 5 most popular hashtags in two different cities: New York and San Francisco.
3
We can consider that New York is inside a box, in wich the bottom-left point has the
coordinates (-74.2557, 40.4957) and the top-right point has the coordinates (-73.6895,
40.9176). San Francisco can also be considered inside a box, with bottom-left point as
(-122.5155, 37.7038) and top-right point as (-122.3247, 37.8545).
For this problem, use the dataset http://vgc.poly.edu/projects/gx5003-fall2014/
week3/lab/data/sample_data_problem_8.txt. Output the 5 most popular hashtags of New York, with their count, and then do the same for San Francisco.
The hashtags must be sorted by their count value. If the count of two hashtags are the
same, then they should be sorted in lexicographical order.
> p yt h o n p r o blem 8 . py s am pl e _ d at a _ p r o bl em _ 8 . t x t
New York :
# nyc , 119
# HighCla s sHoodlum s , 86
#NYC , 72
# Job , 40
# newyork , 30
San F r a n c i s c o :
# S a nF r a n ci s c o , 17
# s a n f r a n c i s c o , 14
# Job , 13
# Jobs , 10
# sf , 9
The output of the sample is available at http://vgc.poly.edu/projects/
gx5003-fall2014/week3/lab/data/sample_output_problem_8.txt.
Questions
Any questions should be sent to the teaching staff (Instructor Role and Teaching Assistant Role) through the NYU Classes system.
How to submit your assignment?
Your assignment should be submitted using the NYU Classes system. Create a zip
file with your source code problem1.py through problem8.py. Name the zip file as
NetID_assignment_3.zip, changing NetID to your NYU Net ID.
Grading
The grading is going to be done by a series of tests and manual inspection when required. Make sure your code runs on the sample datasets as specified to minimize the
need for manual inspection of the code, which can be very subjective.
4
