A1. Customer Problem Identification- Individual Report [CLO1, CLO2, CLO5]
Length: Max 900 words Description
This assessment provides the opportunity to identify customer problems from product review data and generate new product or service ideas.
After customers purchase and use products or services, they share their experience on online product review platforms. Many companies try to identify customer problems and their unmet needs from large- scale product review data by using natural language processing such as topic modelling and sentiment analysis. While the recent advancementsin AI methods allow us to categorize textdata automatically by doing topic modelling and labelling, these AI methods are not perfect. Thus, your task is to identify customer problems using both an AI machine and manual labelling by humans to generate new product or service ideas.
Details
Task 1: Product Review Categorization
1.1. By AI:
The file A1 Intent.csv contains “recommended intents generated by IBM AI” (See A1 Data Description of AI Intent.docx). Note that we gave 1000 sentences to IBM AI but some irrelevant sentences were dropped by IBM AI.
(a) What is the number of intents generated by the AI Machine?
(b) What is the number of review sentences selected by the AI machine for each intent and across all the intents?
1.2.By human:
To address the ambiguity in some intents generated by IBM AI and an excessively large number of intents relative to the limited review sentences, please do human categorization further. In other words, you can change some intents to different categories. Candidate groups are product (e.g., skincare), product attributes (e.g., longevity), different stages of customer journeys (e.g., online shopping, browsing, delivery), and soon.
(c) How many intents (by AI) did you change to different categories?
(d) What is the number of the final categories labelled by a human (you)? Ensure that it does not exceed 30 categories.
(e) During human categorization, why did you change the initial categories the AI machine had generated? In other words, what kind of categories did you try to make? And why?
Task 2: Importance-Satisfaction Plot
Using the final categories generated from Task 1 and the Python code covered in the lecture and tutorial, please count the intent frequencies, measure the intent sentiment and make an Importance-Satisfaction plot. Then, please interpretate the plot by dividing the plot into four quadrants: Right-Bottom, Left-Bottom, Right-Top, and Left-Top.
(a) What does each quadrant mean?
(b) What categories are located in each quadrant?
Task 3: Customer Problem Identification & New Product or Service Idea Generation
Based on the importance-Satisfaction plot, please make recommendations for the company.
(a) what are the company doing relatively well?
(b) what are the primary customer problems?
(c) What does the company need to improve with high and low priority?
(d) To address the identified customer problems, suggest new product or service ideas.
When completing the tasks above, please apply appropriate data analytics practices and integrate key concepts introduced in class. Ensure that your discussion is logical, clearly structured, and professionally presented. Your report should not exceed the word limit, excluding the title page, relevant images, tables or charts.
Title page (1 page) includes (1) the Title of your report, (2) the Word count, (3) the Course name, tutorial session and group, tutor’s name, (4) Your first and last name & zID.
Submission instructions
A. Submit your report to Turnitin via Moodle.
- .doc contains your report. Filename: “Tutorial session_Group_ your first and lastname & zID _A1.doc” (e.g., T9 1 Junbum Kwon_zXXXXX_A1.doc)
B. Submit other supporting files (data, and code) to Moodle submission folder.
1) .xlsx file contains data including Human categorization columns which you added.
2) .ipynb contains all relevant code to get the results in your report.
● For each missing file among the above (1) to (2), -1 mark
Marking Criteria
Your assignment will be marked based on the following marking criteria:
1. Analysis: Quality of analysis - categorization and plotting
2. Interpretation & Recommendations: Quality of interpretation and new product idea
3. Written Presentation: Quality of written report For further information, see the below marking rubric.