Skip to content

README-data.md

Latest commit

 

History

History
69 lines (50 loc) · 3.91 KB

README-data.md

File metadata and controls

69 lines (50 loc) · 3.91 KB

README.md for data

This "data" folder contains two sub-folders:

1) "XBRLs": XBRL-files of company's annual or quarterly reports. Currently, these files do NOT contain ESG-data.
2) "JSONs": JSON-files of company's sample ESG-data that later will be used to populate a NEO4J-Knowledge-Graph

I. XBRLs:

A) Where to get the files from?
Companies today already publish their annual or quarterly reports in XBRL-format, but these XBRL-files do not yet contain ESG-data. From 2024 on, this will change as ESRS requires companies whose size exceeds certain thresholds to publish ESG-data in the form of XBRL-files.
To show how such XBRL-files can be automatically parsed and read into JSON-files, here three sample XBRL-data-packages are provided for the following companies: Adidas AG ("adidasag.com"), BASF AG ("basfse.com") and Philips ("www.philips.com").
These XBRL-data-packages were downloaded from: ESMA Databases and Registers - Corporate Reporting

B) Files Format
The downloaded XBRL-data-packages must be organized in a certain manner so that the Python script "A_read_xbrl.py" later can parse and read these XBRLs into a JSON-file properly. The downloaded XBRL-files usually come in the form/hierarchy of:
- META-INF
- reports
- url-of-company
    - xbrl
        - report-year
            - file-name.xsd
            - file-name_cal.xml
            - file-name_def.xml
            - file-name_lab.xml
            - file-name_pre.xml
These files must be reorganized as follows:
- raw
    -> FOR ALL COMPANIES, COPY ALL "url-of-company"-FOLDERS (AND FILES THEREIN) INTO THIS "raw"-FOLDER:
    - url-of-company-1
    - url-of-company-2

    - reports 
        -> FOR ALL COMPANIES, COPY ALL XHTML-FILES IN THE ORIGINAL "reports"-FOLDER INTO THIS "reports"-FOLDER:
        - file-name.xhtml

    -> LEAVE THE OTHER FOLDERS (SUCH AS "www.esma.europa.eu", "efrag.org", "www.xbrl.org", "xbrl.ifrs.org") 
       UNTOUCHED AS THEY ARE NEEDED FOR PARSING THE FILES.
The XBRL-files now can be parsed and read into JSON-files with the Python script "A_read_xbrl.py".
If there are some problems while parsing, please refer to the attached document "XBRLs/XBRL_explained.pdf" or the GitHub page of the Python package that was used for parsing: py-xbrl
If an "xbrl.TaxonomyNotFound"-error occurs, please check the raised issue here: py-xbrl issues

II. JSONs:

As we wanted to inject some sample ESG data into the Knowledge-Graph, we created some exemplary JSON-files assuming that such files in the course of the year 2024 will be extractable from the to-be-reported XBRL-files (see above). The required ESG data to be reported was outlined by ESMA and published in an Excel table: ESRS_Draft_10_2023.xlsx
For readability reasons and clarity, we did not want to inject sample data for all to-be-reported data points into the Knowledge-Graph, but only for some of them. We used the following 21 data points:

The "JSONs"-sub-folder contains the following exemplary JSON-files containing these data points for three companies ("Adidas", "BASF", "Puma") and two years (2022, 2023), altogether 6 JSON-files:
- Adidas_2022.json
- Adidas_2023.json
- BASF_2022.json
- BASF_2023.json
- Puma_2022.json
- Puma_2023.json
To create such exemplary json-files to be later read into the Knowledge-Graph, the template "templates/template.json" can be used.

The next section is: Models