An Analysis Of Worldwide Catalog Data, 2018/2019 Version

Well this is embarrassing...

I was checking the traffic to my website, and I saw that I had been linked a while back by Bob Skinner at Filling Spaces. He had posted back in May about the update mentioned above.

On his post, a commenter had opined that the numbers appeared low for classic-era (1840-1940) stamps. So, I went back to the source data and sure enough, there was a major omission in the data.

The original source data was stored in about 500 different database files, and I wrote some code to import these hundreds of individual databases into a master database, then compile all the data into one massive table/spreadsheet. Somehow, when the import process was running, it skipped a handful of files, including a large chunk of Europe. I didn't notice because the end result was a table with almost 600,000 rows, and it was impossible to check the accuracy of all of the data.

I went back in and re-ran the data import and updated the website.

A few interesting stats:

1840-1940 total: 99,766 issues
1840-1940 total mint CV: 31,978,784 €
1840-1940 total used CV: 19,206,986 €

1840-2018/2019 total: 759,032 issues
1840-2018/2019 total mint CV: 40,309,423 €
1840-2018/2019 total used CV: 24,864,878 €

In addition, Michel lists many stamps without a date. Some of these are unissued stamps. Not sure why there's no date on the remainder. To see these you have to set the dates as 0 (zero). This adds an additional:

4,545 issues
mint CV: 972,800 €
used CV: 528,678 €

These totals are far higher than the commenter noted above estimated for the classic era. This is likely the result of Michel's listings (especially Germany) generally being more specialized than Scott or SG -- see the "Detailed IAs" version of the data to get an idea of how deep Michel goes in many areas. There's also still some degree of duplication due to the way Michel catalogs things (ex. Michel lists souvenir sheets/blocks and the individual stamps).

But anyway. The data is a little closer to accurate, so feel free to check it out again. It'll never be 100% because I don't have the time or inclination to double-check 763,577 entries.