Dataset: Predicting Profitable Customer Segments

About this Dataset

Predicting Profitable Customer Segments

Context

Marketing is a key component of every modern business. Companies continuously re-invest large cuts of their profits for marketing purposes, trying to target groups of customers who have the potential to bring back the highest Return On Investment for the company. The cost of marketing can be very high though, meaning that the decision about which customer group to target is of great financial importance.

This dataset was made available by an online retail company that has collected historical data about such groups of customers, tracked the profitability of each individual group after the respective marketing campaign and retrospectively assessed whether investing on marketing spend for that group was a good choice.

Content

In order to enable machine learning experimentation, this dataset has been structured as follows:

Each row is a comparison between two groups of potential customers:

Column names starting with "g1_" represent characteristics of the first customer group (these were known before the campaign was run)
Column names starting with "g2_" represent characteristics of the second customer group (these were known before the campaign was run)
Column names starting with "c_" are features representing some comparison of the two groups (also known before the campaign was run)

The last column, named "target", is categorical, with 3 categories:
0 - none of the two groups were profitable
1 - group1 turned out to be more profitable
2 - group2 turned out to be more profitable

Inspiration

Can you build a machine learning classifier that accurately predicts which of the 2 groups (if any) will turn out to be more profitable?

Tables

Customertargeting

@kaggle.tsiaras_predicting_profitable_customer_segments.customertargeting

1.46 MB
6620 rows
71 columns


CREATE TABLE customertargeting (
  "g1_1" DOUBLE,
  "g1_2" BIGINT,
  "g1_3" BIGINT,
  "g1_4" BIGINT,
  "g1_5" BIGINT,
  "g1_6" BIGINT,
  "g1_7" BIGINT,
  "g1_8" BIGINT,
  "g1_9" BIGINT,
  "g1_10" BIGINT,
  "g1_11" BIGINT,
  "g1_12" BIGINT,
  "g1_13" DOUBLE,
  "g1_14" DOUBLE,
  "g1_15" DOUBLE,
  "g1_16" DOUBLE,
  "g1_17" DOUBLE,
  "g1_18" DOUBLE,
  "g1_19" DOUBLE,
  "g1_20" DOUBLE,
  "g1_21" DOUBLE,
  "g2_1" DOUBLE,
  "g2_2" BIGINT,
  "g2_3" BIGINT,
  "g2_4" BIGINT,
  "g2_5" BIGINT,
  "g2_6" BIGINT,
  "g2_7" BIGINT,
  "g2_8" BIGINT,
  "g2_9" BIGINT,
  "g2_10" BIGINT,
  "g2_11" BIGINT,
  "g2_12" BIGINT,
  "g2_13" DOUBLE,
  "g2_14" DOUBLE,
  "g2_15" DOUBLE,
  "g2_16" DOUBLE,
  "g2_17" DOUBLE,
  "g2_18" DOUBLE,
  "g2_19" DOUBLE,
  "g2_20" DOUBLE,
  "g2_21" DOUBLE,
  "c_1" DOUBLE,
  "c_2" BIGINT,
  "c_3" BIGINT,
  "c_4" BIGINT,
  "c_5" BIGINT,
  "c_6" BIGINT,
  "c_7" BIGINT,
  "c_8" BIGINT,
  "c_9" DOUBLE,
  "c_10" BIGINT,
  "c_11" BIGINT,
  "c_12" BIGINT,
  "c_13" BIGINT,
  "c_14" BIGINT,
  "c_15" BIGINT,
  "c_16" BIGINT,
  "c_17" BIGINT,
  "c_18" BIGINT,
  "c_19" BIGINT,
  "c_20" DOUBLE,
  "c_21" DOUBLE,
  "c_22" DOUBLE,
  "c_23" DOUBLE,
  "c_24" BIGINT,
  "c_25" DOUBLE,
  "c_26" DOUBLE,
  "c_27" DOUBLE,
  "c_28" DOUBLE,
  "target" BIGINT
);