FastCompany has a great article this week on the results of the 1870 census, and the hand-made graphics (“BI –2.0”?) that were made from the data (thanks to RadicalCartography.net and the Library of Congress). [NOTE -- the original version of this post erroneously said "1830" throughout]

Here’s a selection of my favorites – click on each one to get the full graphic.

US geology – the hand-shading is so much nicer to look at than computer-generated graphics:

Nationality data mapped:

Here’s a chart that would have looked good on a 19th century Recovery.org website (where, incidentally, data visualization guru Edward Tufte is now a Presidential Advisor). There’s a lot of history illustrated here:

• The light blue spike in the lower right is the US Civil War (1861-1865), which prompted the introduction of a “temporary” progressive income tax (the light pink blob in the lower left) that was originally supposed to expire in 1866!
• The pink blobs on the right are a record of expenditures on the US Indian wars
• The dark blue spike in the middle of the left-hand chart shows land sales in the West – which may have caused the collapse of the banking system in 1837

This is a beautiful rendering of church-going in the US. Interestingly, rather than show non-churchgoers as a separate bar, they are relegated to a grey box around the outside.

Here’s a nice example of showing data through proportions.

And here’s proof that men are bigger idiots than women! (of course, they meant what we’d call “mentally ill” these days)

Population charts showing that the “cowboy states” (e.g. Wyoming, Nevada, Montana) were populated by young male foreigners, while Utah had LOTS of children.

This chart is a good example of something that hasn’t changed much in the last couple of centuries – the notion that you should show data just because you have it. For example, Chart 1 below show deaths per month per state. Is this data actionable in any way?

Later charts aren’t nearly as attractive, like these, from the 2000 Census:

Here’s a PowerPoint file with some of the key charts already included:

If you liked this post, you might also like:

• Giant Brains: Did Edmund Berkeley Predict Email, the Internet, ERP Systems and the iPhone in 1949?
• The Real Pioneer of Business Intelligence (and BI 2.0)?
• Performance Excellence, Past and Future?

I wrote about this site and others in a post in July, Great Examples of US Government BI Transparency. Unfortunately, it turns out that it’s not so great, in that some of the data shown was completely false… (this is generally considered a reasonable baseline of things to avoid).

Seth gave the example of the graphic on the left below: “The Green and Purple segments are approximately equally sized even though the green segment should be 25% larger than the purple segment”.  Thankfully the problem has now been corrected, and the site now shows what the chart should look like (with the data values now slightly updated) – see how much larger the “grants” area is now.

The problem turned out to be with the way the site calls the Google Chart API. The values passed to the engine for display were absolute values, not percentages, so they are being truncated. (I’m biased, of course, but to me, that sounds like a good reason to rely on a full BI solution, rather than the Google Chart API… )

The real big cause for concern is that it was apparently only when the webmasters read Seth’s blog posting that they realized they had a problem…