Telling stories with misguided numbers

How to tell different stories with the same numbers:
[Via Only in it for the gold]

Here is a figure from Reuters about US employment numbers. The worst is over, huh?

Here is a figure from the New York Times showing essentially the same data. Hmmm. Notice how the longer time scale, and the expression of the total rather than the month-over-month change, changes the picture substantially and gives a much clearer picture of what is going on.

One of Edward Tufte’s main points is that the way you display data affects the lesson people take from the data; that aligning the facts with human psychology such that people extract the real picture is a deep skill.

Of course, in the case of our friends who are arguing that global warming has stopped, you could equally argue that deliberately misaligning the facts with human psychology is also a deep skill.

Numbers are used all the time to tell a narrative. One must be very careful in analyzing numbers by graphing them. One is to make sure the complete time period is show. As with climate change, there is often a trend but with a high noise level. Thus, individual years may be cooler, due to noise, but the overall trend is hotter.

Cherry-picking the data can be misleading. Why did Reuters only show the data from the beginning of 2008 and why did it only show the change from each month? Also why look at total number change rather than percentage, which would reflect how the job markets change?

Probably so that you can only take away from the story exactly what they decide to tell you. There is really very little further interpretation possible.

The second graph is much more informative. First it does show just how bad 2008 is but we can also compare it to other periods. In the previous recession, the job markets continued to drop[ for quite some time after the recession ended.

We are already below that bottom and the recession is not yet over. Just how bad is it going to get? Plenty bad.

Tufte is great for examining the narratives that can be created with graphs. If one wants to hide the truth, graphs can be used to obfuscate. So be wary of graphs, particularly from people with an ax to grind. Make sure they are not playing fast and loose with the facts

Technorati Tags: General, Science

Animals for drugs

by Mclaire2
US advisers back 1st drug from DNA-altered animals – Reuters:
[Via Google News]

ROCKVILLE, Maryland (Reuters) – The first drug made using genetically engineered animals to near U.S. approval won key support on Friday from an advisory panel that judged it safe and effective despite concerns from groups worried about the genetic tinkering.

GTC Biotherapeutics Inc’s experimental anticlotting therapy, called Atryn, is made using a human protein gathered from female goats bred to produce it in their milk.

[More]

Producing therapeutic drugs in animals rather than in human cell culture or bacteria could have substantial effects on the costs associated with many drugs. The physical plants needed to produce many protein-based drugs are extremely costly. Actual capacity is limited and can take years to build.

In addition, there are difficulties with the cost of manufacture itself (i.e. the specialized media needed to grow the cells) as well as waste disposal problems after the protein has been produced.

Pharming overcome these difficulties with a more sustainable approach. Not only are media costs much less (i.e. potentially actual chicken feed) but the production levels (that is, the amount of protein per volume of ‘media’) can be much greater. This means a herd of 150 goats could produce the world’s supply of a drug, rather than a $100 million manufacturing facility.

Using farm animals hold advantages over using plants. Plant fertilization can be problematic, since pollen is air-borne. It is much easier to monitor the whereabouts of animals than of plants, so it is unlikely that the GM animals will trade genes with other animals. Production animals can be neutered so that no genes will be passed in any unintended fashion. Disposal could proceed much like current drug facilities but the ‘waste’ should be much smaller. Energy costs would be less.

Animals from rabbits to cows can be used. The economic footprint of this approach appears to be much lower than the huge footprint of a 10,000 liter manufacturing plant. In addition, scaleup would be much easier and more quickly performed using animals.

Often new products render a current therapeutic obsolete. Currently, the expensive manufacturing facility must then be modified for a new product or sold to someone who needs the capacity. Mothballing a facility can be very expensive and modifying it can also entail a large cost, which is passed onto us. Not so with animals.

A big plus is that instead of only producing drugs with a billion dollar market, the economies of this approach could make it easier to produce drugs for much smaller markets. Thus a more sustainable approach for the production of therapeutics could be maintained.

Technorati Tags: Environment, Science, Sustainability

Tips for any web site

by Noël Zia Lee
Using Social Media Efficiently: 52 Tips from Beth Kanter:
[Via TechSoup Blog blogs]

Wow. Beth Kanter has impressed the heck out of me again. She’s participating in Convio’s Now is the Time campaign of New Year’s resolutions for nonprofits and technology (along with our very own Robert Weiner who’s own resolution post is here.)

Beth’s New Year’s resolution is to use social media efficiently and she offers up 52 great tips on how to do so. And if you try one each week for the next year, maybe we can all be better communicators, have more effective and engaging campaigns, and mobilize our supporters for greater change. New Year’s resolutions always make me dreamy.

Some of my favorites from her list include:

Do an annual ROI for your blog (and other social media activities) using benchmarking and metrics.
Don’t set up a presence on every social network in the world all at once.
If you are not reading blogs and Web sites in an RSS Reader, make that your New Year’s resolution.

read more

These rules are not only for non-profits but are important for anyone who is using social media to connect people or move information around. Many of these can be adapted to either non-profit or for-profit situations.

The key is capturing the right metrics. Web 2.0 approaches create a treasure trove of data that can be effectively mined to learn just what is working and what is failing, leading to effective solutions. So do not only have a plan for getting Web 2.0 tools up and going but also have a plan for mining the data they produce.

Technorati Tags: Open Access, Web 2.0