Charting Charts Chart


From Priceonomics which notes:

In our sample of Times pieces, we only found one chart used outside the Business section before 1990….By the late 2000s, charts had become central to the journalism at the Times. The newspaper had a team of graphics reporters infusing the paper with innovative visualizations. For the Times web edition, those data visualizations were often interactive.

…The New York Times cemented their prioritization of visual journalism when they launched the data-driven news site the Upshot in 2014. The Upshot’s original team of fifteen included three full-time graphic journalists.

Computers are the obvious driving factor. Data analysis and visualization are powerful tools. I would guess that data IQ has been increasing ala the Flynn effect.


I remember when everyone was wringing their hands over how horrible "USA Today" was. Now it's the model for most news.

No. Al Jazeera and the BBC are the models of journalism these days. Everyone else is playing he said / she said clap-trap for ratings

The demonstrative no! Very good, if he were saying what you think he's saying.

Shame, because the BBC is utter garbage and getting worse by the day.

USA Today started in 1982, just at the inflection point of that chart. I wonder if there is a relationship?

Of course this is it.

Even as Paul Romer, the new chief economist at the World Bank, questions the accuracy of data-driven macro.

Propagandists understand the power of visualization.

"a picture is worth a thousand words"

A rare rayward comment I am in total agreement with.

How does this track with Global Warming and Pirate Attacks?

And how does it track with NYT's paid circulation numbers?

I don't often find correlation proves causation convincing. But when I do I choose nonstationary time series.

While the New Yorker has a deservedly high reputation it assumes a math literacy level that is pre 1500. When was the last time you saw a chart or graph in the New Yorker?

I only read it for the pictures.

By that time, newspaper readers (office workers), started to use spreadsheet software with chart-making capabilities: Lotus 1-2-3, MS Excel etc It would be curious to find a letter to the Editor from that time asking for charts.

Geez, like any other institution, change came slowly to the NYT. At one time there were no photographs at all, now there are color photos. Are you people advocating not using visuals to explain a story to maintain tradition or what? This is a dopey post.

I would think that it would be an indication of the progress of printing technology. Not so long ago a chart would have needed to be drawn, photographed, then transferred to the printing plates. At one point that no longer was the case, and if it is easy, it gets done.

Oh, I thought it was the USA debt chart.

If you like lots and lots of charts and graphs, you might check out this compilation of mine (in Kindle ebook form):

Because charts = information!

I read NY Times articles on Japan and some of the econ charts and graphs have been off or misleading, although common mistakes journalists make. Worse were a few charts and graphics on radiation from the Fukusihima accident.

I'd love to see the "one chart used outside the Business section before 1990"

found in "a" September issue checked every 5 years. That's a total sample size of 32. My math ability isn't the greatest: could someone walk me through the statistics showing that an exponential model is a better fit than a flat (zero) line before 1935 and a linear fit after that? Does this graph correlate with the proportion of the adult (ie. paid subscriptions) public who were exposed to graphing (mathematical functions) in (high) school? I gotta admit, I don't recall ever seeing a pre-WW2 algebra (or possibly trigonometry) text book... but I'd think graphs and algebra would be taught together. Charts? IDK.

I dunno, looks like "Peak Charts" may be happening...

In the 1970s I worked in a couple of large government economic research organizations.

If you wanted to a chart in your report you had to calculate all the data points by hand and give them to a graphic artist in another division. You would discuss it with the graphic artist who would proceed to produce the chart. From the time the process started till the final chart was approved by all those involved normally took a week and that is assuming that some error or other problem did not come up during the entire process.

To do a regression we used punch cards and the computer division would run it overnight and you would see the results the next day. The punch cards included the instruction on how the computer was to process the data. If so, a correction was needed a series of punch cards would have to be added. We kept the punch cards in a cardboard box. In case their was an accident and the box was dropped we numbered each punch card with a black marker. Whenever we made a correction to the program the new cards would be given a letter and numeric value. So maybe if a correction had to be made at, for example card number 350 you would retain the originally numbered cards but starting at 350 you would create a series of cards numbered 350A to ???A at which point you directed the computer to return to some point in the originally numbered cards. If you made an error you would not know until the next day. So, just running a simple regression would normally take a week or more to finish.
And this was considered state of the art.

The original large scale econometric models were computed using a mechanical computer
( adding machine) and the people doing it would make hundreds if not thousands of runs on the mechanical computer or adding machine.

What is the statistics / math education of the people creating these charts? Or is PowerPoint / excel experience the only requirement?

Based on what I've seen with international news, no education in stats, econ or physics (Fukushima) but there are exceptions. Rogoff once said on NPR that foreign journalists know nothing about economics (NY Times, WaPo, etc) but that domestic economics coverage is often quite good, and he wasn't sure why. One reason is that foreign correspondents are area studies and history majors with no econ background yet sometimes write econ stories.

"The Visual Display of Quantitative Information" by Edward Tufte was published in 1983 and pretty quickly reached classic status.

Tufte mentioned in it how much further ahead Japanese culture was at quantitative graphical literacy at the mass level. In the 1990s, I noticed that Pokemon cards sometimes used sophisticated graphics that I'd otherwise only seen in "The Economist."

+1. Great book.

We should always keep eye on news and all these sort of stuff whether it’s from here or anywhere else. I do just that and thanks to broker like OctaFX, I find it extremely easier with their daily market news and analysis service, it’s deadly accurate so following it makes a valid difference and is something that allows me to be entirely at ease and comfort with doing my job and it’s what makes me very much relaxed with the work.

Comments for this post are closed