Export Board Metadata, Natural Language, Tags, etc

  • 29 November 2020
  • 8 replies
  • 976 views

I am trying to better analyze a very complex board formed from the:

https://wiki.p2pfoundation.net/Main_Page


Because of it’s complexity I was hoping to access its objects from this python client (which wraps around the REST API)

https://github.com/jasonlarkin/miro-python-client/blob/develop/gameb.ipynb

https://github.com/jasonlarkin/miro-python-client/blob/develop/miro/client.py

I noticed this topic came up recently:

https://community.miro.com/got-a-question-ask-the-community-45/export-pictures-and-full-board-from-miro-2258

I don’t necessarily need to export pictures, just the natural language associated with any metadata would be helpful.  As of now this metadata is missing (see gameb.ipynb)

 

Having this metadata could enable semantic analysis of the board which could compliment the same analysis on this wiki which was used to seed the board:

https://wiki.p2pfoundation.net/Main_Page

My plan is to perform analysis in this vein:

https://www.mdpi.com/1999-5903/8/2/14/htm

and publish the results. 
Gracias!

Jason

 


8 replies

Userlevel 3
Badge +7

Hello Jason,

 

The analysis that you want to do looks very interesting. Are you an academic researcher?


Regarding your question about natural language, we are not planning on adding any extra features to the SDK in the near future, as we are currently focusing on improving the current functionality. Having said that, could you elaborate a bit more on what metadata you expect to be able to retrieve? In particular, could you explain what you mean with “the natural language associated with any metadata would be helpful”?

 

Kind regards.

 

Daniela

Developer Experience Team

Userlevel 3
Badge +7

Hello Jason,

 

Sorry for the delayed response. Thank you for sending the detailed information about your research. Could you clarify what you mean with “natural language statements contained in the shapes/widgets”? Do you mean the text inside the widgets, like for example “Spirituality” and “Philosophy” in the green STICKER widgets in your last screenshot? If that is what you are referring to, it is possible to retrieve that through the API and SDK. 

About specific user data, that would not be possible due to privacy concerns.

About data as a function of time, we do not have plans at the moment for such a feature.

 

Kind regards.

 

Daniela

Badge

Hi Jason - 

 

I was blown away when I found your miro board that I found on this site- https://spaceshipearth.live/ 

I would love to connect with you about what you are working on the Noosphere that you created in Miro, and beyond. I ran an experiment on Clubhouse www.eventforhumanity.com where we spent 4 weeks crowdsourcing best practices for Regional Resilience, using Miro & discord Otter and Raindrop to capture what we discovered. 

Anyway- if you are still reading this thread let me know and maybe we can connect in the future. 

Cheers

TYLER

 

 

 

“The analysis that you want to do looks very interesting. Are you an academic researcher?”

Yes, I am pursuing this analysis as part of a “knowledge management system” research team formed out of:

https://www.complexityweekend.com/

my intent is to analyze the large miro board with at least these tools (and likely more, such as NLP/NLU):

https://www.mdpi.com/1999-5903/8/2/14/htm

“Regarding your question about natural language, we are not planning on adding any extra features to the SDK in the near future, as we are currently focusing on improving the current functionality. Having said that, could you elaborate a bit more on what metadata you expect to be able to retrieve? In particular, could you explain what you mean with “the natural language associated with any metadata would be helpful”?”

The metadata that would be useful would be:

  1. The labels, natural language statements contained in the shapes/widgets of this board. Specific user data would also be helpful, in the vein of https://www.mdpi.com/1999-5903/8/2/14/htm
  1. These same metadata, but as a function of time (in the vein of https://www.mdpi.com/1999-5903/8/2/14/htm)

These, plus additional NLP/NLU tools (which I don’t have a compete list of as yet, but probably many of the same tools available in https://iris.ai/) would enable me to study the miro board, much as wikipedia was studied in https://www.mdpi.com/1999-5903/8/2/14/htm.  I am working on getting access to: https://wiki.p2pfoundation.net/Main_Page, where I will be attempting the same type of analysis.  The miro board in question is related (somewhat) to the p2p wiki as the users/topics of each have some shared.

Essentially, I am looking for metadata that would enable analysis that is not possible via visual inspection/interaction as the current system allows.  The full topical extent of this board is enormous, it covers topics that span a huge chunk of civilization/history/ontologies/etc

 

I could give you access to the board, here are some samples

 

here’s close to a full zoom-out

Here are some local shots

 

Whatcha think?

 Could you clarify what you mean with “natural language statements contained in the shapes/widgets”? Do you mean the text inside the widgets, like for example “Spirituality” and “Philosophy” in the green STICKER widgets in your last screenshot? If that is what you are referring to, it is possible to retrieve that through the API and SDK.

 

 

yup that’s exactly what I mean.  I was able to access part of the meta data from the python REST API wrapper on GitHub 

https://github.com/jasonlarkin/miro-python-client/blob/develop/gameb.ipynb

https://github.com/jasonlarkin/miro-python-client/blob/develop/miro/client.py
Do you know how to access the rest (ie Spirituality, etc)?

 

on user data/anonymity yea that’s totally fine, would be it still be possible to get an anonymous user-id (example “user-39baj8b7vv701&)? Just a label to differentiate contributions/interactions

 

on time, the board admin Johan C Branstedt can provide me with snapshots of the board at different times, so I can use that for any temporal analysis.

 

thanks for the help!

 

Userlevel 3
Badge +7

HI Jason,

Stickers have a property called “text” where you can find the text written in the sticker:
https://developers.miro.com/reference#sticker
 

Looking at your code, you would have to check when the widget type is “STICKER” and make sure you fetch the “text” property from the json object.

https://github.com/jasonlarkin/miro-python-client/blob/develop/miro/utils.py#L41

 

The text on STICKER widgets can contain html, so you might want to remove the html tags (usually header tags).

 

You can also check all the properties of a particular STICKER if you have its id if you paste the url on your browser:

https://api.miro.com/v1/boards/<BOARD_ID>/widgets/<WIDGET_ID>?access_token=<YOUR_ACCESS_TOKEN>
 

 

I hope this helps.

 

Kind regards.

 

Daniela

Reply