# Beginners time series predictions — Data

来源：https://www.kaggle.com/competitions/beginners-time-series-predictions/data

本地 Meta Kaggle 快照；不是本次实时页面。

# Dataset Description

## Intro

I am monitoring some files 24/7 via an automated monitoring system. Of all the monitored files, I selected 2 with 2 datapoints each (size & modification time -- see more below).

The target is to predict the next values (see `sample_submission.csv`/`test.csv`).

## Files

*   **train.csv** - the training set
*   **test.csv** - the test set
*   **sample_submission.csv** - a sample submission file in the correct format

## Columns

*   `Id` - combination of "itemid" and "clock" - please see below.
*   `value` - the value

## `Id` interpretation

The `Id` column can be splitted into 2 columns like this:
```python
df[['itemid', 'clock']] = df.Id.str.split('_', expand=True).astype(int)
df = df[['itemid', 'clock', 'value']]
```

The `itemid` consists of these unique values:
* 36182: size of file1
* 36185: modification time of file1
* 36188: size of file2
* 36191: modification time of file2

The `clock` column is a UTC unix timestamp when the observation has `been made. fo

## Good to know

The data is for Belgium, so you might want to incorporate Belgium holidays in your logic as well.