CUSTOMER SERVICE ANALYTICS q

Learning Goal: I’m working on a python writing question and need an explanation and answer to help me learn.

create atoolthat allows the user to assess the linguistic properties of

major companies’ customer service tweets. Your program should first ask the user to specify
the type of analysis they would like to perform, and it should then ask the user which company
they would like to analyze. Your program should allow the user to perform an unlimited
number of analyses.

The dataset is available at https://dgoldberg.sdsu.edu/515/customer_service_tw…
and contains approximately 1,000 tweets for each of twelve major companies’ customer service
Twitter handles, listed alphabetically below:

@amazonhelp @AppleSupport @Ask_Spectrum @AskPlayStation
@comcastcares @hulu_support @SpotifyCares @sprintcare
@TMobileHelp @Uber_Support @UPSHelp @XboxSupport

The data is formatted as a JSON array. An example of the formatting is below:

[ {“Company”: “@sprintcare”, “Text”: “I understand. I would
like to assist you. We would need to get you into a private
secured link to further assist.”},
…
{“Company”: “@UPSHelp”, “Text”: “Hello, please click the
link to let us know how we can assist you. Click the link
to DM us with your tracking and phone number. ^E.W
https://t.co/wKJHDXWGRQ”} ]

Your tool should support the following types of analyses:

• Polarity: your tool should calculate the average sentiment polarity across that
company’s customer service tweets. Ensure that you calculate sentiment scores per–
tweet rather than per–sentence.

• Subjectivity: your tool should calculate the average sentiment subjectivity across that
company’s customer service tweets. Ensure that you calculate sentiment scores per–
tweet rather than per–sentence.

• Formality index: this metric is an attempt to estimate how formally or informally text is
written based upon its parts of speech (find the part of speech tagging scheme here).
Scores close to 0 reflect very informal text, and scores close to 100 reflect very formal

text. Your tool should calculate the average formality across that company’s customer
service tweets. Formality is computed as follows:

𝑓 = 𝑛𝑢𝑚𝑏𝑒𝑟 𝑜𝑓 𝑛𝑜𝑢𝑛𝑠,𝑎𝑑𝑗𝑒𝑐𝑡𝑖𝑣𝑒𝑠,𝑝𝑟𝑒𝑝𝑜𝑠𝑖𝑡𝑖𝑜𝑛𝑠,𝑎𝑛𝑑 𝑑𝑒𝑡𝑒𝑟𝑚𝑖𝑛𝑒𝑟𝑠
𝑐 =𝑛𝑢𝑚𝑏𝑒𝑟 𝑜𝑓 𝑝𝑟𝑜𝑛𝑜𝑢𝑛𝑠,𝑣𝑒𝑟𝑏𝑠,𝑎𝑑𝑣𝑒𝑟𝑏𝑠,𝑎𝑛𝑑 𝑖𝑛𝑡𝑒𝑟𝑗𝑒𝑐𝑡𝑖𝑜𝑛𝑠

50( (𝑓−𝑐)/(𝑓+𝑐 )+1)

For the purposes of this assignment, treat any tag that contains “NN” as a noun; any tag
that contains “JJ” as an adjective; any tag that contains “IN” as a proposition; any tag
that contains “DT” as a determiner; any tag that contains “PR” as a pronoun; any tag
that contains “VB” as a verb; any tag that contains “RB” as an adverb; and any tag that
contains “UH” as an interjection.

Write your code such that the entirety of your program is case insensitive (for example, the
program should behave equivalently if the user enters “yes”, “Yes”, or “YES” or if they enter
“@sprintcare” or “@SprintCare”). However, do not perform any spellchecking on this
assignment (unfortunately, the spellchecker we have used corrects “polarity” to “popularity”).

Some considerations as you write your code:

• Consider the possibility that, when loading the dataset, some connection issue occurs
(that is, a status code other than 200). Ensure that your code handles this case and
provides the user with a helpful printout if it does occur.

• Consider the possibility that the user enters an unsupported mode (something other
than polarity, subjectivity, or formality). In this case, print out a short warning message
and allow the user to continue analyzing data. It is also possible that the user will enter
an unsupported Twitter handle, but you do not need to handle this case for this
assignment.

• Ensure that your prompts and output are crisp, professional, and well–formatted. For
example, ensure that you have used spaces appropriately and checked your spelling.

• Adding comments in your code is encouraged. You may decide how best to comment
your code. At minimum, please use a comment at the start of your code to describe its
basic functionality.

Please use the following as a template for the tool’s expected functionality:

Welcome to the customer service analyzer!

Which analysis would you like to perform
(polarity/subjectivity/formality)? POLarity

Which Twitter handle would you like to analyze? @upshelp

@upshelp: –0.020923390993824426

Would you like to run another analysis (yes/no)? yes

Which analysis would you like to perform
(polarity/subjectivity/formality)? Subjectivity

Which Twitter handle would you like to analyze? @SPOTIFYCARES

@SPOTIFYCARES: 0.29245743199261076

Would you like to run another analysis (yes/no)? yes

Which analysis would you like to perform
(polarity/subjectivity/formality)? Formality

Which Twitter handle would you like to analyze? @xboxsupport
@xboxsupport: 59.37499613549362

Which analysis would you like to perform
(polarity/subjectivity/formality)? finances

Which Twitter handle would you like to analyze? @amazonhelp

Sorry, that type of analysis is not supported. Please try again.

Would you like to run another analysis (yes/no)? no