How to Find Accurate and Compelling Data

No matter your message, including data to support your claims enhances the validity and impact of whatever you are communicating -- but only if your data comes from a reliable source. Too often, individuals cherry pick data from a variety of ... let’s say “questionable” sources, cobbling together a few statistics that are incomplete -- or worse, completely irrelevant.

Whether you’re communicating to your boss or your consumer, all data starts with a good data source.In this post, I'll go over what a good data source looks like. Then, I'll list a few of my favorite places to look for reliable data and research.

好数据源的组成部分

一个好的数据源是...

Original

无论您是使用内部数据还是外部数据，请确保它始终来自主要来源。如果您在Wikipedia或杂志文章中遇到有趣的数据集，请跟踪原始数据。如果不亲自查看主要数据，您将永远不会知道这些方法是否存在缺陷，样本量太小，或者问卷调查表现出偏见。

和伙计们...朋友不让朋友引用维基百科。Remember: It's your responsibility to represent the data in as close to its original context as possible.

综合的

如果您对数据没有提出的疑问，您的受众也会。您的数据源应提供足够的信息来获取大局以及任何适当的上下文。例如，假设您的销售指标显示在4月至5月之间出售的单位有很大的跳跃 - 还不错！

But looking at the larger set, you see that jump came after a general decline. That’s the full story you want your data to tell.

Current

The world changes quickly. Oftentimes, data that was produced or collected ten years, five years, or even a year ago is obsolete. Always use the most recently published version of the data available. Some data producers, such as the Department of Labor, revise their data on an annual or even monthly basis -- but this is not always the case for every data producer.

Ideally, you'll want your data to be one or two years old at most. Beyond this, use discretion; and in all cases, be upfront about how old the data set is. For example, the graphic below uses data that's a few years old, but they made sure to specify that it's the most recent year available in the bottom lefthand corner.

Reliable

验证您选择的来源是否相关，合法且尽可能无偏见。强大的资料包括政府机构收集/生产的数据，例如美国人口普查局或劳工部汇编的统计数据。其他顶级数据来源可以包括行业白皮书或学术出版物。

请记住，投票机构或智囊团进行的调查虽然有用，但通常会有政治议程。因此，与老化数据的情况一样，使用酌处权。避免使用轶事，用户生成或基于传闻的数据。

判断来源是否合适，请问自己以下几点：

• Who wrote this web page? Does the author have credentials?
• 这个网页是否隶属于可信组织？
• 网站上次更新何时更新？
• 托管网站的组织的目的是什么？
• Does the author provide a bibliography?

Where to Find Good Data

幸运的是，有许多可靠的来源可以使您可以访问有关各种主题的大量数据。这里有一些可以让您入门。

General Data/Research

• UNDATA: A statistical database of all UN data.
• 亚马逊公共数据集：与生物学，化学，经济学等有关的大数据集的存储库，包括人类基因组项目。
• 皮尤研究：公众舆论民意测验，人口研究，媒体研究等。

学术研究/白皮书

• 谷歌学术: A wide array of information, including articles, theses, books, abstracts, white papers, court opinions,etc.

财务数据

• Google Finance: Real-time stock market data.
• Google公共数据资源管理器: Searchable large datasets on economic development worldwide.

政府/世界数据

• 中央情报局世界概况：有关历史，人，政府，经济，地理，通讯，交通，军事等的全球信息。
• U.S. Census Bureau: Demographics, geographic information, and education for U.S. population.
• data.gov: An open data source for U.S. government data.

Social Data

• 社会管理：实时社交数据。
• Google趋势：探索搜索数据和趋势。
• Topsy: Search social trends and tweet history.

健康数据

• healthdata.gov：有关医疗补助，医疗保险，clincial研究，治疗等的数据。
• cdc.gov：来自疾病控制与预防中心的公共卫生数据。

只有当您拥有从固体来源的良好数据时，您才能制作出可信，信息丰富且有说服力的信息。

Where else do you go to find good data?