Digital Financial Reporting

The following is a summary of the quality measurements of XBRL-based US GAAP financial reports submitted to the SEC as of September 30, 2018.  The following Excel spreadsheets and other documents provide details related to these quality measurements:

• Negative results from tests (i.e. confirmed errors): Details of 100% of the confirmed errors. 987 errors contained with 655 filings.
• Other information related to testing of US GAAP reports: Lots of details relating to testing of US GAAP reports.
• Other information related to testing of IFRS reports: Lots of details relating to testing of IFRS reports.
• Prior quarter's results: Same measurements for the prior quarter ending June 30, 2018.
• US GAAP reporting styles: Overview of US GAAP reporting styles.
• IFRS reporting styles: Overview of IFRS reporting styles.

US GAAP fundamental accounting concept relations continuity cross check validation results for last 10-K or 10-Q filed by generator of the report as of September 30, 2018: 

(click image for larger view)

US GAAP fundamental accounting concept relations continuity cross check by logical accounting relation tested (same filings as above):

(click image for larger view and more descriptive information about test)

US GAAP fundamental accounting concept relations continuity cross check comparison across periods:

(click image for larger view)

If you want additional information please contact me directly.

**********************PRIOR RESULTS**********************

Previous fundamental accounting concept relations consistency results reported: June 30, 2018; March 31, 2018; November 30, 2017; August 31, 2017; May 31, 2017; March 31, 2017; November 28, 2016; August 31, 2016; June 30, 2016; March 31, 2016; February 29, 2016; January 31, 2016; December 31, 3015; November 30, 2015; October 31, 2015; September 30, 2015; August 31, 2015; July 31, 2015; June 30, 2015; May 29, 2015; April 1, 2015; November 29, 2014.

An engineer at Wolfram was kind enough to create a demonstration for me that explained some of the capabilities of the Wolfram | One platform.  Here is that demonstration which you can download and instructions how to get this to run.

The instructions are in the ZIP file.

I hope to create more of these!

I have created a set of GitHub repositories for examples, templates, rules and other metadata related to XBRL-based financial reporting.  Metadata will be created and maintained for US GAAP and IFRS. (I am also creating metadata for a sample reporting scheme that am using for debugging, testing, education, and training I call XASB.)

I also provided a bunch of Excel-based information extraction and validation tools for US GAAP and IFRS.

All XBRL instances, XBRL taxonomy schemas, XBRL linkbases, XBRL formulas have been validated using numerous XBRL processors. The quality should be very high.

All of these examples, templates, rules, and other metadata are still going to be available in their original locations; I am pretty sure that I will be using this more modern GitHub approach to providing this information going forward.  I have some details to work out, but things are looking really good so far.

You might want to grab a tool such as the local GitHub Desktop client application so that you can download all these files to your local computer.  Doing this will create local copies of all the files in the repository on your local computer.

If you want to contribute files you should get an XBRL editor, XML editor, text editor, or code editor; preferably one that has built it Git.  For example, Visual Studio Code has built in Git. The text editor Atom does also.  (What would be nice is if an XBRL taxonomy editor supported Git.)

I will add all the metadata that I currently have to these GitHub repositories, then I will figure out what the next steps are.  I hope people find this useful.  If you have metadata, tools, examples, etc. that you might want to contribute, please contact me and we can get your contributions uploaded.

I wrote the paper Computer Empathy to explain, as best as I could, that in order to get a computer to do what you want you have to understand computers and how to get computers to perform work reliably, predictably, repeatedly, and safely.

As it turns out there is another term that conveys the same message: computational thinking.  There is actually a Center for Computational Thinking at Carnegie Mellon University.  The Center for Computational Thinking and Stephan Wolfram, in his blog post How to Teach Computational Thinking, provide good definitions of computational thinking.  I synthasized those definitions and other ideas into my own definition of computational thinking:

Computational Thinking is a thought process involved in formulating problems and their solutions so that the solutions are represented in a logical, clear, and systematic form that can be explained to and effectively carried out by a computer or human.

This five minute video provides an excellent explanation of what computational thinking is and why it is important. (Here is a playlist of 9 videos if you really want to dive in.)  This is an easy to read three page paper, Conputational Thinking written by Jeannette Wing. And finally, here is a 40 minute video that explains Computational Thinking.

Computational thinking is a process, a technique (set of skills and habits), a philosophy. Computational thinking combines aspects of logic, mathematics, engineering, systems theory, and computer science. Computational thinking helps you differentiate: what humans can do better than computers; what computers can do better than humans. Computational thinking is about leveraging computers as a labor saving device through automation of work.

Computational thinking is made up of four elements:

• Decomposition: breaking a problem down into easy to manage parts.
• Pattern recognition: spotting what different problems have in common and using what has worked before to help you solve a new problem.
• Abstraction: focusing on the details that matter and ignorinig details that don't matter.
• Algorithmic thinking: generate simple steps can be used to solve a problem.

Computational thinking is not just for computer scientists; it is basically for everyone.  Computational thinking is the new literacy for the 21st century.  If you are doing accounting, reporting, auditing, and/or analysis in a digital environment and you are not computational thinking enabled you may not remain relevant.

Computational thinking is about fundamental principles and concepts, not about technology.  Computer Empathy provides principles and concepts; but also digs into the details of how to get computers to perform work effectively.  Others call for essentially the same thing using different terms.  Some call for teaching logic to every high school student. Others use the term critical thinking.

##################################

Course in Computational Thinking for Educators

Computational Thinking, 10 Years Later

Computational Thinking Benefits Society

Wolfram can be used for free.  I don't understand the exact limitations, but there is an browser-based Open Cloud version of Wolfram.  (I would encourage you to use Google Chrome as I was having some issues with Microsoft Internet Explorer.)

There is a very helpful Elementary Introduction to Wolfram that is provided.  That seems like a good place to start.

There is extremely good documentation provided.

Why might you want to learn about Wolfram?  The blog post How to Teach Computational Thinking provides a good explanation.

To understand the sorts of things you can construct with Wolfram, check out Wolfram Alpha. Here are some money and finance related stuff.

Clearly you have to use your imagination a bit.  But try and imagine what accounting, reporting, auditing, and analysis in a digital environment might look like.

If you create anything interesting in Wolfram please let me know.