Use this for a complete guide to do data science in marketing for analysis, from regression, to text analysis.
Collection of Models and Algorithm to solve marketing and sales problem. Some of the functions here are plug-and-play, thus ready to use (see ready-to-use_function folder). Most of the models here are built based on Data Science approach, thus the aim is more to make a prediction rather than to understand causality. But some of them also include the tools to explore explainability of a machine learning model. The models here include:
- Regression
- Looking for the best model for regression
- RFE (Recursive Feature Elimination) to know the most important variables
- Customer Segmentation
- K-Means segmentation
- Optimizing number of segments using elbow method etc.
- Product Recommendation (Cosine Similarity)
- User-based collaborative filtering
- Item-based collaborative filtering
- Text Analytics
- Preprocessing the text using spacy
- Text classification and explanation
- Unsupervised topic modelling
- Text summarization
- Sentiment analysis
- Word embeddings for semantic relationship
- RFM Analysis (rfm_analysis)
- Recency, Frequency, and Monetary segmentation of customers
- Using simple method by dividing data into quintiles
- And a more sophisticated algorithm using KMeans
- Included with a custom-wrapped Python module for doing the analysis easier
- Cohort Analysis (cohort_analysis)
- Categorizing customers based on a shared characteristic or event
- Using a simple method by calculating cohort index
- Then going with a more sophisticated way using Survival Analysis
- Included with a custom-wrapper Python module for doing the cohort analysis using cohort index
All the data set used for the example of these were obtained from Kaggle data set collection. The files are not listed here but instead, it is listed on each of the project file to make the citation of the data set more clear. The link to download the data set is also provided within the project.