What we are supposed to do

1. import data for clickstreams and orders
   

2. clean imported data

• set appropriate variable types
• mark NA values
• delete empty columns (only NAs)

3. create plots
   

• simple univariate descriptive plots
• time series
• grouped plots for comparisons
• more specific plots, e.g., distribution of revenue across products or product categories (long tail), Lorentz-curve

4. further analysis; some ideas

• try to merge clickstreams and orders
• customer segmentation (recency, frequency, monetary value)
• product clustering
• streams of customers between product clusters (sankes diagram)
• sequential patterns in clickstreams and orders

Regarding Programming in R and Python

All exercises should be done in R and Python. For the project, the following rules apply:

Step	R or Python
Import, manipulate	both, R: tidyverse, mostly dplyr , Python: NumPy, Pandas
Plots	up to the teams, but the packages should implement the Grammer of Graphics (R: ggplot2, Python: plotnine)
Documentation, Table	up to the teams
Inference	up to the teams
Prediction	up to the teams (probably better in Python)