TE
科技回声
首页24小时热榜最新最佳问答展示工作
GitHubTwitter
首页

科技回声

基于 Next.js 构建的科技新闻平台,提供全球科技新闻和讨论内容。

GitHubTwitter

首页

首页最新最佳问答展示工作

资源链接

HackerNews API原版 HackerNewsNext.js

© 2025 科技回声. 版权所有。

How to Lie with Data Visualization

120 点作者 matm大约 11 年前

8 条评论

j2kun大约 11 年前
It&#x27;s interesting to see the y-axis scaling example, because the author uses it to sell as incomplete a truth as the document he&#x27;s criticizing.<p>&gt; displaying the data with a zero-baseline y-axis tells a more accurate picture, where interest rates are staying static.<p>This is not true of all data, or even interest rates. Who&#x27;s to say that a 0.01 increase in interest rate <i>doesn&#x27;t</i> translate to large differences in money? If this is true (and it depends a lot on the intended audience for the chart), then representing it with a zero baseline is far more deceptive and the &quot;bad&quot; chart does a better job showing that the small increases are in fact getting steadily bigger.<p>The problem is that data alone is not sufficient to say anything, and the context (in this case, noise in interest rate changes) is too important to give a single rule for your scale. How you interpret context, of course, depends on what you&#x27;re selling.
评论 #7589458 未加载
评论 #7589374 未加载
Mithaldu大约 11 年前
The last example is especially striking to me, as i&#x27;ve seen that graph used in a discussion about the effects of gun laws. The person posting it had taken the time to remove the numbers.
评论 #7589282 未加载
nathancahill大约 11 年前
I highly recommend reading &quot;How to Lie with Statistics&quot;[0]. Old but still very relevant.<p>[0] <a href="http://www.amazon.com/How-Lie-Statistics-Darrell-Huff/dp/0393310728" rel="nofollow">http:&#x2F;&#x2F;www.amazon.com&#x2F;How-Lie-Statistics-Darrell-Huff&#x2F;dp&#x2F;039...</a>
评论 #7589165 未加载
评论 #7589266 未加载
评论 #7589159 未加载
andremendes大约 11 年前
Great article, I&#x27;ve seen those techniques being used many times by brazilian media. It&#x27;s not easy to spot those lies at a glance but it&#x27;s a must if you don&#x27;t want to get fooled by traditional media news.
评论 #7589237 未加载
jmharvey大约 11 年前
The article doesn&#x27;t even mention the most egregious offender on display in their examples: the tilted&#x2F;3-D pie chart. In the Romney&#x2F;Palin&#x2F;Huckabee graphic, Palin&#x27;s wedge has the smallest area, despite having the largest number.
评论 #7589357 未加载
MarcScott大约 11 年前
This reminds me of an episode of The Saturday Night Armistice from way back in 1995<p><a href="https://www.youtube.com/watch?feature=player_detailpage&amp;v=686tttWspnU#t=771" rel="nofollow">https:&#x2F;&#x2F;www.youtube.com&#x2F;watch?feature=player_detailpage&amp;v=68...</a>
ASneakyFox大约 11 年前
What about when people make charts where the x and y axises are unrelated or stupid? If you plot oxygen consumption and death you&#x27;d find that almost all people who died last year consumed oxygen.
dsugarman大约 11 年前
a logarithmic or exponential scale could be effective in lying as well