Shih-Chun Liu sl3497
1. How to use this program.
With your tweets text file, 'zips.csv', and 'sentiments.csv' in the same
folder as 'tweeter.py', run the 'tweeter.py' module, or import it in another
program to run in the Python Shell and use the functions written in the
code. To test the module, you may run 'tester.py' in the Python Shell and
follow the instructions.
2. Design decisions.
This section of the module continues to operate on the tweets list of tweet
dictionaries from assignment 6A.
I split the assignment into the following functions:
-make_sentiments
This iterates through every line of 'sentiments.csv', saving into the
sentiments dictionary the word as a key and the float as the
sentiment value for that key.
-remove_Nones
This function comes in useful for calculating the average value in a
list, since sum(list) will not work if None is in the list.
-average
I wrote this because a lot of these sentiment functions need to
calculate the average of a list of sentiments.
-calc_sentiment
This uses tweets_words from 6A to create a list of the words in a
tweet, then appends each word's sentiment into a separate list of
sentiments, and plugs this list into the average function to find
and return the tweet sentiment.
-add_sentiments
This performs make_sentiments, then iterates through the tweets list
and adds the sentiment to each tweet after using calc_sentiment.
-tweet_sentiment
Like the one-liner functions from 6A, this returns the sentiment of
a tweet input.
-write_newtweets
Same as write_tweets from 6A, with sentiments added.
-tweet_filter
If any of the keys' value in kwargs is an empty string, this function
replaces that value with None.
This first creates a copy of the tweets list. Then it checks whether
each key is in kwargs and whether the value associated with that key
is None, then creates a new list and appends the appropriate tweets
to that list, leaving out the tweets that are filtered out. It
continues to filter the new list until finished, and then returns the
new filtered list. For word and state, I made sure to lowercase the
word and uppercase the state before filtering, just in case the user
is not consistent with the expected input.
-avg_sentiment
This saves the sentiment of each tweet into a new list, then plugs
that new list into average and returns that average tweets sentiment.
-make_state_sentiments
This function comes in useful for the extra credit functions. It
iterates through tweets, adding each state as a dictionary key and
the sentiments of tweets from that state into a list whose key is
the state. Then it returns the state:[sentiments] dictionary.
-most_positive
This first filters tweets for the given word, then uses
make_state_sentiments to create a state:[sentiments] dictionary for
that filtered list. Then it sets an initial sentiment of None and
iterates through each list in that dictionary, calculating and saving
the highest average sentiment and state of that highest average
sentiment. If there are only sentiment values of None, then it sets
the most positive state to None. Then it returns that state.
-most_negative
This works the same as most_positive; however, I used an initial value
of 2 instead of None because all floats have greater values than None.
3. Conclusions.
While operations with word sentiments are practical, what I have done is
limited to the 'sentiments.csv' file. This program would be more useful if I
could also add and update the 'sentiments.csv' file to account for words that
do not have sentiments saved in that file. Also, another limitation is
singular vs plural words--they show up as different words in the way I have
coded. Furthermore, **kwargs opens up many more creative ways to write
functions.