文档章节

气候数据只能预测降雨么?那你就错了!

openfea
 openfea
发布于 2017/03/02 10:49
字数 1510
阅读 10
收藏 0

“转自:灯塔大数据;微信:DTbigdata”

——————————————————————————————————————————————————

天气状况与人类息息相关,不管你身处在地球的哪个角落,都无法忽视天气和气候给人类生活造成的影响。全球各个地区的气候有着巨大的差异,然而生活在某个固定区域的人可能不会明显感觉到这种差异。

天气和气候状况也是影响一个地区经济的重要因素,气候一旦出现异常,经济就可能会受到严重的负面影响,比如洪涝灾害会损坏重要的基础设施,而长期不下雨则会导致干旱的发生。

气候对社会来说影响巨大。有人甚至认为叙利亚内战就是由天气变化所导致的,美国前任国务卿约翰·克里说过:“气候变化不一定是叙利亚危机的直接导火索,但是很明显,叙利亚发生的旱灾是火上浇油、雪上加霜。”

因此,关注天气状况并不是仅仅为了知道明天应该穿什么,而是需要公司和企业制定一个长期的计划,根据天气数据判断未来几周的天气状况会对其业务造成什么样的影响。

众所周知,英国的连锁超市乐购(Tesco)就利用气候数据预测销售情况,调整仓库库存。举例来说,如果他们接到天气预报数据说接下来三周会有明显升温,他们就会检查仓库中是否有足够的烤肉架、防晒霜和冷饮等来满足消费者的需求。

乐购2013年的数据就显示,当年凭借天气预测数据,该集团节约了超过600万英镑,将脱销成本降低了30%。

但是2013年之后的几年,全球气候变得更加变幻莫测,2014年、2015年、2016年是有记录以来全球气温最高的几个年份,给人们预测天气增加了很多不确定因素的干扰,给公司营销方案的制定造成了很大困扰。

难以捉摸的气候造成的负面影响包括:损坏基础设施从而导致供应链中断、恶劣天气导致农作物收成差,从而导致某些产品供应短缺。

从这种意义上来说,气候数据不仅仅能够帮助人们预测灾害的发生,还能帮助人们制定合理的方案,以减少突发气候灾害对公司业务所造成的负面影响。

影响气候的数据点数量庞大,想要挨个输入这些数据点,然后推测出某个地区出现不同天气得可能性,这无疑是个异常艰巨的任务,需要消耗无数的人力和能源。

美国国家海洋和大气局(NOAA)投入巨资打造克雷(Cray)超级计算机,这种超级计算机的计算速度高达每秒3千万亿次。

要将如此多影响气候变化的因素考虑在内,就必须依靠这种强大的计算力,越来越多的公司和企业也将会依赖于这些气候数据来帮助他们提高经营的效率。

众所周知,在公司经营中,外部影响因素和内部主观能动同样重要,天气和气候就是外部因素的其中一方面。天气和气候对每个公司都有一定程度的影响,因此我们有必要好好的利用气候数据来预测天气状况,并制定出高效的应对方案。

 

英文原文

Weather data isn’t just about predicting rain

The weather is perhaps the one universal that every human can relate to, it is the universal connector between everyone regardless of where you are from. Most people have seen broadly the same kinds of weather, despite geographical differences meaning that others see more extreme versions than others.

It is also one of the most important elements of any economy, after all, with any kind of excess of one type of weather, there is likely to be significant disruption, whether this is excessive precipitation destroying vital infrastructure or over abundance of dry weather causing droughts. It also has a profound impact on society as a whole, with some even saying that weather has played a part in starting the Syrian civil war, with John Kerry, then Secretary of State saying ‘I’m not telling you that the crisis in Syria was caused by climate change, but the devastating drought clearly made a bad situation a lot worse.’

It is therefore not simply a case of looking at the weather forecast to ascertain what to wear that day, companies need to be able to use long-term planning to predict how the weather in several weeks may impact them.

UK based supermarket chain Tesco have become well known for their use of weather data to help predict sales and stock requirements. For instance, if they can predict that there will be a heatwave in 3 weeks time, they can then make sure that they have increased stocks of disposable barbecues, sun cream and cold drinks to cater for the increased demand. When this was widely reported in 2013 it was said that the chain had managed to save £6m ($7.5m) per year and seen a cutting out-of-stock by 30% on special offers.

However, since 2013, when Tesco’s use of this data became more publicly known, the world’s weather has become considerably more complex. 2014, 2015, and 2016 have all been the hottest years on record, which has caused more unpredictability in our weather and more disruption to companies. This can be anything from supply chains being disrupted due to infrastructure damage through to shortage of stock because farmers couldn’t grow a specific crop due to difficult conditions. It means that the use of data is essential not only in the prediction of these events, but in the planning in case they occur.

There are a huge and diverse number of data points that can impact changes in the weather and trying to input every one and then predict with any degree of certainty down to localities is incredibly difficult and requires considerable power. It is why the NOAA invested millions in a Cray supercomputer that processes 3 quadrillion calculations per second. The huge variations in conditions that go into weather conditions requires this kind of power and is something that more and more companies are needing to look at to help run their businesses effectively.

Everybody knows that running a business effectively is as much about awareness of external influences as it is about what you do internally, and weather is the ultimate example of that. Every company can be impacted and the use of data to predict and react to it is essential.

 

翻译:灯塔大数据

本文转载自:

openfea
粉丝 18
博文 86
码字总数 95615
作品 1
杭州
其他
私信 提问
Catch Rain If You Can!3.1万美刀召集测雨英雄!

  期待新赛事的天池宝宝们,小天给你们带来了新的赛事消息! 全球气象AI挑战赛已于昨日发布!奖金池有3.1万美金,优胜团队还会获得1,500美金的差旅补贴,去新加坡IEEE ICDM 2018大会参与颁...

天池大数据科研平台
2018/06/06
0
0
香港科技大学施行健:深度学习用于短临降雨预报的一个基准和一个新模型 | 分享总结

雷锋网AI科技评论按:本文介绍了NIPS 2017论文: Deep Learning for Precipitation Nowcasting: A Benchmark and A New Model 中提到的核心算法模型。在近期雷锋网(公众号:雷锋网)举办的GAI...

杨文
2018/01/01
0
0
干货 | 香港科技大学施行健:深度学习如何用于短临降雨预报

  AI科技评论按:本文介绍了 NIPS 2017 论文: Deep Learning for Precipitation Nowcasting: A Benchmark and A New Model 中提到的核心算法模型。在近期举办的GAIR大讲堂线上直播课中,该...

AI科技评论
2018/01/02
0
0
香港科技大学施行健:深度学习用于短临降雨预报的一个基准和一个新模型

  雷锋网AI科技评论按:本文介绍了NIPS 2017论文: Deep Learning for Precipitation Nowcasting: A Benchmark and A New Model 中提到的核心算法模型。在近期雷锋网举办的GAIR大讲堂线上直...

深度学习
2018/01/02
0
0
你不曾见过的酷炫地图可视化作品(二)

猜谜猜谜啦!你一定猜不出最后一张图是什么! 一、前言 本文是《你不曾见过的酷炫地图可视化作品(一)》的系列文章,上回整理了诸如:世界顶级学者全球迁移图、百度春运迁徙图、中国人口分布...

Deserts_X
2017/11/14
0
0

没有更多内容

加载失败,请刷新页面

加载更多

redis 学习2

网站 启动 服务端 启动redis 服务端 在redis 安装目录下 src 里面 ./redis-server & 可以指定 配置文件或者端口 客户端 在 redis 的安装目录里面的 src 里面 ./redis-cli 可以指定 指定 连接...

之渊
昨天
2
0
Spring boot 静态资源访问

0. 两个配置 spring.mvc.static-path-patternspring.resources.static-locations 1. application中需要先行的两个配置项 1.1 spring.mvc.static-path-pattern 这个配置项是告诉springboo......

moon888
昨天
4
0
hash slot(虚拟桶)

在分布式集群中,如何保证相同请求落到相同的机器上,并且后面的集群机器可以尽可能的均分请求,并且当扩容或down机的情况下能对原有集群影响最小。 round robin算法:是把数据mod后直接映射...

李朝强
昨天
4
0
Kafka 原理和实战

本文首发于 vivo互联网技术 微信公众号 https://mp.weixin.qq.com/s/bV8AhqAjQp4a_iXRfobkCQ 作者简介:郑志彬,毕业于华南理工大学计算机科学与技术(双语班)。先后从事过电子商务、开放平...

vivo互联网技术
昨天
24
0
java数据类型

基本类型: 整型:Byte,short,int,long 浮点型:float,double 字符型:char 布尔型:boolean 引用类型: 类类型: 接口类型: 数组类型: Byte 1字节 八位 -128 -------- 127 short 2字节...

audience_1
昨天
11
0

没有更多内容

加载失败,请刷新页面

加载更多

返回顶部
顶部