Subject: Text Mining

Topic selected: Application of text mining on customer satisfaction $($Nike$)$

Data collected: Nike online store reviews from kaggle

Project Background

Customer satisfaction is a critical focus in e$-$commerce, especially for global brands like Nike, which aims to stay competitive, protect its positive image, and continually improve products. With abundant customer reviews, social media comments, and feedback, Nike possesses valuable data that reflects consumer experiences, preferences, and pain points. This project focuses on using text mining to analyze feedback from the Nike Online Store, helping the brand identify trends, sentiment$,$ and issues within customer narratives. Insights gained from this analysis are intended to guide improvements in products and services to enhance the overall online shopping experience.

Business Analytics Concepts$/$Issues

Key text mining concepts in this project include:

Sentiment Analysis: This technique determines the emotional tone of customer feedback, classifying comments as positive, negative, or neutral. By assessing overall sentiment$,$ Nike can quickly identify areas that require attention.

Data Preparation: Preparing data for text mining involves text cleaning, stemming, tokenization, and stop$-$word removal, making the data more suitable for analysis. This step enables Nike to detect patterns and trends that improve customer satisfaction.

Topic Modeling: Topic modeling, particularly Latent Dirichlet Allocation $($LDA$),$ uncovers themes within customer feedback. This segmentation helps Nike address common issues and tailor solutions based on recurring topics.

Project Objectives

Identify Customer Pain Points and Preferences: Analyzing textual feedback allows Nike to identify common issues with the online store, prioritize improvements, and develop strategies to resolve recurring problems.

Enhance the Customer Experience: By understanding what customers appreciate or dislike, Nike can improve website usability, customer support, and product offerings.

Measure and Improve Brand Sentiment: Monitoring sentiment related to specific products or services helps Nike understand customer loyalty and maintain its reputation in a competitive market.

Generate New Marketing and Product Strategies: Text mining insights direct marketing and product development investments, allowing Nike to create tailored recommendations and marketing efforts that align with customer preferences.

Role of Text Mining

Text mining is vital for achieving the project$$s goals, as it extracts insights from unstructured text data. Key processes include:

Parsing and Pattern Recognition: Text parsing breaks down feedback, allowing algorithms to categorize emotions and topics, thereby summarizing overall customer sentiment$.$ This helps Nike prioritize improvements based on dissatisfaction trends and customer preferences.

N$-$gram Analysis: N$-$gram analysis identifies frequent word sequences in feedback, highlighting popular phrases and common issues. These insights inform Nike$$s decisions on enhancing website usability and customer support.

Named Entity Recognition $($NER$)$: NER categorizes proper nouns, helping Nike track mentions of specific products. By analyzing feedback on these items, Nike can gauge brand sentiment$,$ assess loyalty, and develop strategies tailored to market demand

Question: Use the CRISP$-$DM framework to organize the report. IBM SPSS$\text{'}$s text mining is used for this project.