Digital Financial Reporting

I updated a prototype application that I created for extracting information from XBRL-based financial reports submitted to the SEC.  It is an Excel spreadsheet application.  You can download the application which is contained in this ZIP archive.

Just press the button "Compare all on List" on the "Compare" sheet.  What will happen is the application will extract information from 375 financial reports from banks.  It takes about 3 minutes to run depending on you Internet connection speed.

What happens is that information for each bank is loaded into a row in the spreadsheet.  Notice that there are ZERO ERRORS in the data.  The reason for that is that I selected only banks that had no errors in their information.

This is an excellent learning tool if you want to reverse-engineer how to extract information.

Email me if you have any questions.

The graphic below shows the results of my quarterly measurements of fundamental accounting concept relations continuity cross checks of XBRL-based financial reports of public companies which are submitted to the SEC.

I have been measuring specific fundamental accounting information of public company XBRL-based financial reports being submitted to the SEC for going on four years now.  The quality of the information reported continually improves each time measurements are taken, which is a good sign.

Highlights for this quarter include three new filing agents/software vendors that have reached the point where 90% or more of the XBRL-based reports they create are consistent with all of my business rules: Compliance Xpressware, EZ-XBRL, and IRIS Carbon.

You can download a ZIP file that contains an Excel spreadsheet that a list of all 994 errors.

Here is a comparison of March, August, and November 2017.

Information Extraction Prototype: Lets you extract data from 375 XBRL-based financial reports, Download ZIP Archive containing Excel spreadsheet

(Click for larger image)

Here are the errors by test:

  Click image for larger view) 

One tactic that I use to understand something is to compare and contrast what I want to understand to something else.  I compared and contrasted the Semantic Web Stack and the XBRL stack. What became very apparent is the need for a Unifying Logic Framework for Business.

Let me explain.  First, here are some things about what I call the Unifying Logic Framework for Business that you should understand.  The framework itself is for expressing business logic, it is not the business logic itself.  This is what the framework would achieve: 

What would a business professional's interaction with something built using this framework look like? Basically, it would be very similar to interacting with a spreadsheet or a pivot table.  This is achieved using proven software creation techniques, the concept is proven to be feasible.  In the past I have described this as a semantic spreadsheet or NOLAP. The business logic glues things together, not presentation artifacts.

What form would such a Unifying Logic Framework for Business take?  Does such a framework already exist?  What do business professionals lose by not having such a framework?  Who chooses what the framework is from the alphabet soup of ISO/IEC, OMG, W3C, and XBRL International standards? Is the unified logic defined by: 

• ISO/IEC Common Logic (CL)?
• OMG Semantics of Business Vocabulary and Business Rules (SBVR)?
• W3C RDFS + OWL + RIF/SWRL syntax logic? (SWRL is not a recommendation, only a submission, RIF and SWRL seem to have issues)
• W3C RDFS + OWL + SHACL syntax logic which specifies closed world assumption and unique names assumption?
• Industry Initiative RuleLog which is designed to be appropriately expressive for supporting knowledge representation in complex domains and yet to be efficiently implementable?
• Industry Initiative RuleML which allows for partially constrained logic profiles and fully-specified logic semantics?
• XBRL Formula? (which has known deficiencies)

Could a de facto Unifying Logic Framework for Business be sufficient?  More to come...stay tuned.

Well this is really interesting.  Check out the Financial Regulation Ontology. More information here.

The creator of the Financial Regulation Ontology describes it as follows: 

FRO is a foundational part of a Semantic Web approach to regulatory compliance for financial institutions . Ontology Web Language (OWL) is a W3C standard with proven scalability, handling complexity in Bio and Medical field. In the ontology everything is a triple: data, schema, mapping, transformations, rules, … everything is stored a uniform cells of subject-predicate-object.

Yes, this is very interesting.  More to come I am certain.