/NLP-Healthcare-App-Reviews

Analyzing customer needs and deriving design insights from healthcare app reviews using Natural Language Processing (NLP) techniques.

Primary LanguageJupyter Notebook

Mobile Apps Reviews Analysis using LDA Topic Modeling

In this project, I analyzed English mobile app reviews using LDA topic modeling. I collected and preprocessed mobile app review data, built an LDA topic modeling model, determined the number of topics, and derived insights into user needs, preferences, and areas for improvement through visualization of the topic modeling results.

The process of the project was as follows:

  1. Collection of Mobile App Review Data: I collected mobile app review data for analysis in the project.
  2. Data Preprocessing: I preprocessed the collected review data. In this step, I removed special characters, numbers, and stopwords from the reviews and tokenized the words to prepare them in a suitable format for analysis.
  3. Construction of LDA Topic Modeling Model and Determination of Topic Number: I built an LDA model using preprocessed data. This model was used to classify each review into a specific topic and determine the number of topics.
  4. Visualization of Topic Modeling Results: I used visualization tools such as pyLDAvis to visualize the results of topic modeling. This allowed me to understand the frequency of words in each topic and the correlations between topics.
  5. Deriving Insights: Based on the visualization results, I identified user needs, preferences, and areas for improvement. This information can help app developers or marketing professionals in determining the direction for app enhancements.

For the complete code and data related to the project, please refer to the NLP-final.ipynb file.

The repository provides the dataset used for training, the list of stopwords, the list of replacement words, and the visualization results of the topic modeling.