Monday, April 29, 2013

Fighting Woodpeckers of Shellpot Creek

These two woodpeckers were fighting and teasing each other.  I think they were trying to claim the trees in our yard in order to impress lady woodpeckers.  These are red-bellied woodpeckers according to the pictures in the Delaware Museum of Natural History woodpecker collection.  (Cornell lab link with sounds!)

Incidentally taken with my new Panasonic DMC-ZS20 camera.  it is a  point and shoot with a really good 20X optical zoom Leica lens.  They were very high up the tree in the back by Shellpot Creek.  I am pleased with my new camera!

Monday, April 22, 2013

More Excel based destruction - this time in support of austerity

A coding error in excel, among other errors, has seriously weakened the conclusions of an important financial paper (by Harvard economists Carmen Reinhart and Kenneth Rogoff, "Growth in a Time of Debt." ) regarding GDP growth at high debt to GDP levels. Another paper, ("Does High Public Debt Consistently Stifle Economic Growth? A Critique of Reinhart and Rogoff," by Thomas Herndon, Michael Ash, and Robert Pollin) tracked down the flaws - Excel spreadsheet errors, excluded data, and unusual weightings of statistics. With the importance f Excel as a modeling and calculation tool at our company, always be aware of the possibility of errors. 

When modeling and examining model results I especially think of three things among many others - 1. Do the final results make sense? 2. If not, is there an error in my calculations or in the original data? Is there an source external to the model I can check against? 3. If the result is unexpected, is this a new insight? The more surprising, the more evidence needed to convince others, especially in business.

Read more at:

Wednesday, April 10, 2013

Beware the excel spreadsheet - especially when it can destroy the economy!

A variety of links about the ubiquitous use of excel in business and about the dangers of poorly designed software and models. Excel is a great liberator and allows the masses to create many different models, it also allows the freedom create bad or incorrect models and its very ubiquity means that it is likely to end up at the center of some big mistakes.

 In my model creation I try to separate data from calculations in a spreadsheet and I am a big proponent of database style formats (read columns) because then I can run a pivot table or better yet, put the data into Tableau for visualization. But then again much of my personal fun work in excel are modesl and simulations or data manipulation that doesn't need much rigor and will only be used by me so I am free to be creative. I do remember programming in other languages when I was young and in graduate school but now I am stuck with just excel. 

This is a great link and great comment thread by programers who see the threat and promise of excel and how it fits into the bigger picture of creating software to do what the usefual spreadsheet was doing. 

A longer article from The Atlantic 

two more views on bad software from Baseline Scenario.