LaurentiuStancioiu/Cluj-Napoca-Weather-OpenWeather-data

---
license: odbl
task_categories:
- tabular-regression
language:
- en
tags:
- climate
size_categories:
- 100K<n<1M
---

## Cluj Napoca Weather Dataset

A weather dataset of Cluj Napoca taken from the OpenWheather History API. It was scraped from the Open Weather
Map using their Weather API. The data collected was from January 1st 2008 until May 2023
at an hourly rate. The columns are presented in the image below:


### Long Time Series Predictions

Clone the [Time-Series-Library](https://github.com/thuml/Time-Series-Library) repository.
```
git clone https://github.com/thuml/Time-Series-Library.git
```

For training the model create a folder called dataset. After that you can either modify the .sh files (for me it didn't work), or run the [predict.ipynb](https://huggingface.co/datasets/LaurentiuStancioiu/Cluj-Napoca-Weather-OpenWeather-data/blob/main/predict.ipynb)
jupyter notebook.
Alternatively, check the HuggingFace implementations for [Informer](https://huggingface.co/docs/transformers/model_doc/informer),
[Autoformer](https://huggingface.co/docs/transformers/model_doc/autoformer), [Vanilla Transformer](https://huggingface.co/docs/transformers/model_doc/time_series_transformer),
[PatchTST](https://huggingface.co/docs/transformers/model_doc/patchtst) and [PatchTSMixer](https://huggingface.co/docs/transformers/model_doc/patchtsmixer)

### Results

Without any hyperparameter tuning the results on this dataset were the following:


A description of what I did to get to those results can be found [here](https://huggingface.co/datasets/LaurentiuStancioiu/Cluj-Napoca-Weather-OpenWeather-data/blob/main/Long%20Term%20Time%20Series%20Forecasting%20for%20Cluj-Napoca%20Weather%20Prediction.pdf).
Also, the trained checkpoints are [here](https://huggingface.co/datasets/LaurentiuStancioiu/Cluj-Napoca-Weather-OpenWeather-data/tree/main).
The data is licensed per OpenWeather under the [odbl license](https://opendatacommons.org/licenses/odbl/).

原始信息汇总

Cluj Napoca Weather Dataset

概述

• 数据来源：从OpenWeather History API获取的克卢日-纳波卡（Cluj Napoca）的天气数据。
• 数据收集时间：从2008年1月1日至2023年5月，每小时记录一次。
• 数据格式：表格数据，具体列信息如图所示。

数据处理与模型训练

• 时间序列预测：使用Time-Series-Library库进行时间序列预测。
• 训练步骤：
  • 克隆Time-Series-Library库。
  • 创建名为dataset的文件夹。
  • 可以选择修改.sh文件或运行predict.ipynb Jupyter笔记本。
  • 参考HuggingFace上的实现，如Informer、Autoformer、Vanilla Transformer、PatchTST和PatchTSMixer。

结果

• 无超参数调优的结果：具体结果如图所示。
• 详细描述：获取结果的方法描述可在此处找到。
• 训练检查点：训练好的检查点可在此处获取。

许可证

• 数据许可证：根据OpenWeather的odbl许可证进行许可。