Talk:Training data-sets
Verify that this is not a duplicate of the Cynefin exemplar kits article, and, in case, merge the two pages. -----Luca OrlassinoT-A-L-K 16:41, 8 March 2021 (UTC)
- doubt clarified: 2 distinct things. -----Luca OrlassinoT-A-L-K 09:32, 16 March 2021 (UTC)
Have added the standard exemplar set to the exemplar page to make the difference clear --GregBro (talk) 08:28, 9 March 2021 (UTC)
It would be really useful to add next to each Main Label why they were picked, in relation to the three domains. Also an explanation for the use of Category and Sub-Category should be provided, as for now it is not clear why one would need those specifications in the data set. --Corinalupu (talk) 21:08, 19 March 2021 (UTC)
- CorinaLupu, on adding why the item has been picked, I'm open, but tend to disagree. If I correctly understood the "spirit of the game", while there should be a certain balance across the three domains, some kind of ambiguity is also beneficial. If the same item is interpreted as ordered by some and as complex by others, that should be ok. -----Luca OrlassinoT-A-L-K 07:35, 20 March 2021 (UTC)
- LucaOrlassino Yes but here, we are not playing the game, we provide items to people who are going to use them for training. So, for now, it implies first to understand the logic the person that wrote the data item used to pick an item. I think making it easy for trainers would be better. --Corinalupu (talk) 08:40, 20 March 2021 (UTC)
- CorinaLupu, on the explanation of Category and Sub-category, I agree that the one provided in the "Data-set description" section (table: "Data-set table column description") is probably too generic. To give some context, when I was creating new data-sets, I sometimes found it useful to add some attributes to help participants more easily recognize an item across different cultures and levels of education. For example, I may not remember what the "Taj Mahal" is, but if I read "Agra, India", and maybe see a stylized picture, it could be enough for me to trigger my memory and allow me to place it on the board. -----Luca OrlassinoT-A-L-K 07:35, 20 March 2021 (UTC)
- LucaOrlassino Ok, I understand, it's to give some kind of hint or indication. Then, I guess Category can be useful if used to give indications, but maybe Sub-Category is dispensable. What do you think? --Corinalupu (talk) 08:47, 20 March 2021 (UTC)
- CorinaLupu, a possible synthesis of the above discussion could be to repurpose the "Sub-category" column, to accommodate an indication of "likely domain". Before doing so, I would seek the opinion of method practitioners (so far, unfortunately, I am not). -----Luca OrlassinoT-A-L-K 13:39, 20 March 2021 (UTC)
Purpose of data sets (and exemplars)
Picking up on the Exemplar vs Data Set distinction.... what is needed in training scenarios is a pre-generated set of data points that can be used primarily in Cynefin contextualization methods and activities (eg. 3 Points, 4 Tables, Butterfly Stamped). A complete "data pack for a training activity" needs to include the 3 exemplars (currently being explored in the Exemplars page) as well as numerous other items that can be assessed as 'more or less complex'. So for training, I believe that (1) these 2 pages/sets of sets DO need to be connected, as each set needs exemplars, and (2) that some these data sets may not be ideal for this purpose (eg which recipe is more complex than another, or which building?). In fact, connecting exemplars and data sets might highlight this challenge - if you can't identify 3 clear exemplars for your data set, it may not be the right kind of data for this purpose. LucaOrlassino and GregBro please weigh in? --Dglanvill (talk) 08:18, 26 March 2021 (UTC)
- Agreed overall, Dglanvill. On (1) I agree that the connection should be explicit and visible, not just here, but also in Butterfly stamping, for example. Not sure, but maybe we should consider merging the 2 pages? On (2), also re. your page edit summary "Should items be relevant to Cynefin domains? Eg not sure how recipe ingredients would help in a CE training activity?", a method practitioner can certainly have more authority in answering than me. However, allow me just a couple remarks on this. (a) The food-recipe example is just in "idea" status, therefore can be thrashed according to the proposed process if there's a good reason for that. No objection from my side. (b) My understanding - and I might easily be wrong - is that forcing abstraction and ambiguity into the interpretation process of a data-set can be somewhat more beneficial than using clearly pre-categorized items. Of course it can be argued that the principle has been pushed a bit too far in the case of recipes. -----Luca OrlassinoT-A-L-K 10:03, 26 March 2021 (UTC)
- I agree that connection between the item and the examplar is for now not explicit, which doesn't make it easy for someone who is looking for a set to use for training. I suggest to replace "Category" by "Examplar type" or "Domain" and "Sub-category" by "Description", so that we can perceive the nuances with the items 'more or less complex' for instance. What do you think? Dglanvill LucaOrlassino --Corinalupu (talk) 16:08, 28 March 2021 (UTC)
Data set status
Do we know how the data set status is determined? When do we consider the set "used"? Luca GregBro --D.Grudzien (talk) 02:14, 20 March 2021 (UTC)
- D.Grudzien, the idea is: according to the process described in the "Process for training data-set creation" section of the page. The initial idea was to make ready-to-use data-set easily recognizable, as well as those that need further development. Of course improvements are always welcome and should be encouraged. The only principle I would not break is consistency across data-sets. -----Luca OrlassinoT-A-L-K 07:07, 20 March 2021 (UTC)
- Also see "Data-set description" section, where a description of each status is proposed in a bulleted list. -----Luca OrlassinoT-A-L-K 07:13, 20 March 2021 (UTC)
- D.Grudzien this is also a wiki so all content can be regarded as being in a perpetual beta so I would say it when anyone has made use of a set and is happy to provide feedback - have a look at the Butterfly_stamping for the two sets that are currently in use. We also need to add the story about creation of the original set.
Sample data-sets
General remarks
Consider adding dataset "Butterfly Stamping Generic Items" referenced in Butterfly stamping. -----Luca OrlassinoT-A-L-K 16:46, 8 March 2021 (UTC)
- "Generic" data-set included. -----Luca OrlassinoT-A-L-K 09:32, 16 March 2021 (UTC)
We need to consider the format - the original used the Avery 5195 template so they could be printed - do we want to use this to make production simpler? --GregBro (talk) 08:30, 9 March 2021 (UTC)
- I get your point. As long as we keep a consistent fields and caption formats, the wiki tables should be easily exportable to .csv, spreadsheet, etc.. I don't see a way do do the editing and the export in a printable format in one step. Having said that, any suggestion towards simpler production is appreciated. -----Luca OrlassinoT-A-L-K 13:43, 9 March 2021 (UTC)
Having all data-sets in one article, together with the data-set concept description and editing process and guidelines has its practical advantages. Separate referencing of single data-sets from other pages seem unlikely, and, if necessary, direct linking can be done to the appropriate section, if needed. However, separate pages for each data-set should also be considered as an alternative. -----Luca OrlassinoT-A-L-K 17:03, 8 March 2021 (UTC)
The original set was 20 sports and had a label for each of clear, complicated, complex and chaotic and confused. Should we look to map the new datasets in a similar way to ensure we have coverage of the individual domains? --GregBro (talk) 08:33, 9 March 2021 (UTC)
- Adding one column or repurposing an existing one (e.g. sub-category) is not a huge rework, so far. But it would be better to do it before adding further data-set content. I understand that this is just to ensure the balance between the domains, not for label printing, right? -----Luca OrlassinoT-A-L-K 13:50, 9 March 2021 (UTC)
Architectural landmarks
Data-set is half-way complete in draft status. All contributions of 5 or more items to the list are more than welcome. -----Luca OrlassinoT-A-L-K 17:11, 8 March 2021 (UTC)
- If minimum requirement of at least 20 items is adequate, then we can consider the data-set ready for testing. -----Luca OrlassinoT-A-L-K 09:32, 16 March 2021 (UTC)
Recipes and ingredients
Data-set is in idea status. Sorting and all contributions of 10 or more items to the list are more than welcome. -----Luca OrlassinoT-A-L-K 17:13, 8 March 2021 (UTC)
Travel destinations
Data-set is in idea status. Sorting and all contributions of 10 or more items to the list are more than welcome. -----Luca OrlassinoT-A-L-K 17:15, 8 March 2021 (UTC)
Corporate functions and departments
Data-set is in idea status. Sorting and all contributions of 10 or more items to the list are more than welcome. -----Luca OrlassinoT-A-L-K 17:16, 8 March 2021 (UTC)
Art work
I have replaced "Category" by "Domain" and "Sub-category" by "Description", as it seemed more pertinent to provide those for people interested in using the data-set. Changes/adjustments in domains and descriptions are welcome. It needs some to increase coherence or provide more clarity. --Corinalupu (talk) 11:50, 7 April 2021 (UTC)