Historical Hourly Weather Data 2012-2017

# Historical Hourly Weather Data

Who amongst us doesn't small talk about the weather every once in a while?
The goal of this dataset is to [elevate this small talk to medium talk][1].

Just kidding, I actually originally decided to collect this dataset in order to demonstrate basic signal processing concepts, such as filtering, Fourier transform, auto-correlation, cross-correlation, etc..., (for a data analysis course I'm currently preparing).
I wanted to demonstrate these concepts on signals that we all have intimate familiarity with and hope that this way these concepts will be better understood than with just made up signals.

The weather is excellent for demonstrating these kinds of concepts as it contains periodic temporal structure with two very different periods (daily and yearly).

![a nice 4 seasons image][2]

### Content

The dataset contains ~5 years of **high temporal resolution** (hourly measurements) data of various weather attributes, such as temperature, humidity, air pressure, etc.
This data is available for 30 US and Canadian Cities, as well as 6 Israeli cities.
I've organized the data according to a common time axis for easy use.
Each attribute has it's own file and is organized such that the rows are the time axis (it's the same time axis for all files), and the columns are the different cities (it's the same city ordering for all files as well).
Additionally, for each city we also have the country, latitude and longitude information in a separate file.

### Acknowledgements

The dataset was aquired using [Weather API][3] on the [OpenWeatherMap website][4], and is available under the [ODbL License][5].

### Inspiration

Weather data is both intrinsically interesting, and also potentially useful when correlated with other types of data.
For example, [Wildfire][6] spread is potentially related to weather conditions, demand for cabs is famously known to be correlated with weather conditions ([here][7], [here][8] and [here][9] you can find NYC cab ride data), and use of city bikes is probably also correlated with weather in interesting ways (check out [this Austin dataset][10], [this SF dataset][11], [this Montreal dataset][12], and [this NYC dataset][13]).
[Traffic][14] is also probably related to weather.
Another potentially interesting source of correlation is between weather and crime. Here are a few crime datasets on kaggle of cities present in this weather dataset: [Chicago][15], [Philadelphia][16], [Los Angeles][17], [Vancouver][18], [Austin][19], [NYC][20]

There are many other potentially interesting connections between everyday life and the weather that we can explore together with the help of this dataset. Have fun!

[1]: https://www.youtube.com/watch?v=qeFlDoepDR0
[2]: http://www.sciencehub4kids.com/wp-content/uploads/2015/08/The-four-seasons.jpg
[3]: https://openweathermap.org/api
[4]: https://openweathermap.org/
[5]: https://opendatacommons.org/licenses/odbl/
[6]: https://www.kaggle.com/rtatman/188-million-us-wildfires
[7]: https://www.kaggle.com/dhimananubhav/2015-nyc-taxi-trips-subset-12-million-rows
[8]: https://www.kaggle.com/kentonnlp/2014-new-york-city-taxi-trips
[9]: https://www.kaggle.com/c/nyc-taxi-trip-duration
[10]: https://www.kaggle.com/jboysen/austin-bike
[11]: https://www.kaggle.com/benhamner/sf-bay-area-bike-share
[12]: https://www.kaggle.com/pablomonleon/montreal-bike-lanes
[13]: https://www.kaggle.com/new-york-city/nyc-east-river-bicycle-crossings
[14]: https://www.kaggle.com/jboysen/us-traffic-2015
[15]: https://www.kaggle.com/currie32/crimes-in-chicago
[16]: https://www.kaggle.com/mchirico/philadelphiacrimedata
[17]: https://www.kaggle.com/cityofLA/crime-in-los-angeles
[18]: https://www.kaggle.com/wosaku/crime-in-vancouver
[19]: https://www.kaggle.com/jboysen/austin-crime
[20]: https://www.kaggle.com/adamschroeder/crimes-new-york-city

## 历史小时级气象数据

于吾辈之中，岂无时或谈及气象之闲谈？本数据集旨在[将此闲谈提升为中谈][1]。

开个玩笑，我最初决定收集此数据集是为了演示基本的信号处理概念，例如滤波、傅里叶变换、自相关、互相关等……（用于我目前正在准备的数据分析课程）。我期望通过对我们熟知的信号进行演示，这些概念将比仅凭虚构信号更容易被理解。

气象数据非常适合演示此类概念，因为它包含具有两个截然不同的周期（日周期和年周期）的周期性时间结构。
![四季之美图][2]

### 内容

本数据集包含约5年的**高时间分辨率**（每小时测量）气象属性数据，如温度、湿度、气压等。
这些数据适用于30个美国和加拿大城市，以及6个以色列城市。
我已经将数据按照共同的时轴进行组织，以便于使用。
每个属性都有其自己的文件，并按时间轴（所有文件中均使用同一时间轴）和不同城市（所有文件中均使用相同的城市顺序）进行组织。
此外，对于每个城市，我们还在单独的文件中提供了国家、纬度和经度信息。

### 致谢

本数据集是通过[Weather API][3]在[OpenWeatherMap网站][4]上获得的，并受[ODbL许可][5]的约束。

### 启示

气象数据不仅内在引人入胜，而且与其他类型数据的相关性也可能具有潜在价值。
例如，[野火][6]的蔓延可能与气象条件有关，出租车需求与气象条件的相关性众所周知（[此处][7]、[此处][8]和[此处][9]可以找到纽约出租车行程数据），城市自行车的使用可能与气象存在有趣的关联（参见[奥斯汀数据集][10]、[旧金山数据集][11]、[蒙特利尔数据集][12]和[纽约数据集][13]）。
[交通][14]也可能与气象有关。
另一个可能有趣的关联来源是气象与犯罪之间的关系。以下是包含在本气象数据集中的城市的一些犯罪数据集：[芝加哥][15]、[费城][16]、[洛杉矶][17]、[温哥华][18]、[奥斯汀][19]、[纽约市][20]。
我们可以利用此数据集探索许多日常生活与气象之间的潜在有趣联系。祝您玩得愉快！

[1]: https://www.youtube.com/watch?v=qeFlDoepDR0
[2]: http://www.sciencehub4kids.com/wp-content/uploads/2015/08/The-four-seasons.jpg
[3]: https://openweathermap.org/api
[4]: https://openweathermap.org/
[5]: https://opendatacommons.org/licenses/odbl/
[6]: https://www.kaggle.com/rtatman/188-million-us-wildfires
[7]: https://www.kaggle.com/dhimananubhav/2015-nyc-taxi-trips-subset-12-million-rows
[8]: https://www.kaggle.com/kentonnlp/2014-new-york-city-taxi-trips
[9]: https://www.kaggle.com/c/nyc-taxi-trip-duration
[10]: https://www.kaggle.com/jboysen/austin-bike
[11]: https://www.kaggle.com/benhamner/sf-bay-area-bike-share
[12]: https://www.kaggle.com/pablomonleon/montreal-bike-lanes
[13]: https://www.kaggle.com/new-york-city/nyc-east-river-bicycle-crossings
[14]: https://www.kaggle.com/jboysen/us-traffic-2015
[15]: https://www.kaggle.com/currie32/crimes-in-chicago
[16]: https://www.kaggle.com/mchirico/philadelphiacrimedata
[17]: https://www.kaggle.com/cityofLA/crime-in-los-angeles
[18]: https://www.kaggle.com/wosaku/crime-in-vancouver
[19]: https://www.kaggle.com/jboysen/austin-crime
[20]: https://www.kaggle.com/adamschroeder/crimes-new-york-city]