# Business Analytics Project Regression — Data

来源：https://www.kaggle.com/competitions/business-analytics-project-regression/data

采集时间：2026-10-19T11:22:27Z

### Data Overview

The data was stored in four separate folders: action, courier, distance, order. Each folder contains a dataframe, labeled by date. Here's some brief overview of some of the attributes, the rest can be explored on your own.

#### Action

The ActionStep attribute records the delivery-man's action, you may group by wave based on courier_id and wave_index. the expect_time is the timestamp when the action was completed, and can be sorted to give the courier's sequence of actions. 'courier_wave_start_lng' and 'courier_wave_start_lat' records the geographic locations. 

#### Distance

'source_type' and 'target_type' are categorical values containing three values: 'Assign', 'PickFood', and 'DeliverFood'. When type is 'Assign', the location attributes represent the courier's location; When type is 'PickFood' the location attributes represent the vendor's location; When type is 'DeliverFood' the location attributes represent the customer's location.

#### Order

| attribute     | explanation                      |

| ------------- | -------------------------------- |

| create_time   | The time the order was created   |

| confirm_time  | The time the order was confirmed |

| assigned_time | The time the order was assigned  |

| aoi_id        | Unique ID indicating an area     |

| shop_id       | Unique ID indicating a vendor    |

### Prediction

For each 'action', group by courier and wave, and check 'expect_time' attribute. The non-zero values are the actual timestamps of completed actions. Your job is to predict the timestamps of zero values in this column. Ideally, after you've predicted all timestamps, ordering by the timestamps gives you the chronological order of the actions.

The available data is from 2.1 to 2.28, and we want you to predict for 2.29