Identify Swedish schools owl:sameas

[salgo60]

There is a register of Swedish schools with a unique id see SCB Skolenhetsregistret

Shouldnt that id be part of the metadata in the specification like owl:same as as many of the open meals will be for schools

• more thoughts

[salgo60]

OT: created a WD object for the specification Q112566739 please change/update it if its wrong

[wincentek]

OT: created a WD object for the specification Q112566739 please change/update it if its wrong

Nice!

Projektet startade på dataportalen, men har därefter flyttat till sambruk (som har ägarskapet). Jag tänker att inget längre skall referera till dataportalen? Dvs, allt skall leda till https://github.com/Sambruk/Open-Meal

Eller?

[bjornhagstrom]

The reason we did not include skolid is that the municipalities do not include this information in the system so we don't have the data unfortunately. It would be good to have.

[salgo60]

@bjornhagstrom

The reason we did not include skolid is that the municipalities do not include this information in the system so we don't have the data unfortunately. It would be good to have.

But that is no excuse.... the data is not useful if you dont have 5 star data and same as - cleaning the data should not be done by every consumer see video

See how Humlab is working with "Riksdagens corpus" this is the first time in 6 years with Wikidata I see people working professional

• we have same as between that data and Wikidata etc... all the corpus is converted and use TEI.... they release the data and have quality control tests

[bjornhagstrom]

I disagree. We can only do so much in getting data owners to create data the themselves do not need or use.

[salgo60]

@bjornhagstrom But then you work in the wrong way what I think we see now is that old organisations as Riksarkivet, Kungliga Biblioteket dont have the skills so to get something done ML researchers creates the data themself... my comment about the quality of Riksdagens data

• issues I see with Riksdagens data

What you say is what I call #77-3 no data is produced and #77-4 Strings not Things its a waste of time and money....

• priducing a spec with 15 fields is done by chatGPT in 10 minutes

[bjornhagstrom]

Yes I am aware of your stance on these things and I agree that you have identified the goals to strive for. But it is not possible to reach them with the first efforts within an area and I do what is possible ot what is perfect. In that regard we differ. With your approach nothing would ever get done since most people are not knowledgable enough to do that. I prefer to build knowledge and understanding for what is possible.

[salgo60]

I prefer to build knowledge and understanding for what is possible.

Give me examples, good examples.... what I see

• 2012 Forum Öppna data
• 2013 E-delegationen
• 2018 Vinnova project to produce Open meal video
• 2022 NSÖD 16 million outcome no data - video has been moved - video new location

I believe in yesterdays weather so 2032 we will still have no data....

• maybe the game changer is chatGPT it gets ridiculous to run projects in 10 months that is produced in 10 sec with chatGPT
  • compare pdf and the same with chatGPT produced in 10 sec
    • and as always we see no data just a pdf.... compare how a dashboard could look
    • same but POC showing Swedish municipals

[bjornhagstrom]

You want us to work differently so it is up to you to say how we can jump directly to all linked data and prove that it is possible across the board.

[salgo60]

• learn from the Humlab project they have very skilled people

Example of team members

• Måns Magnusson Uppsala Universitet assistant professor in Statistics at the Department of Statistics, Uppsala University
• Fredrik Noren working as a senior research assistant at Humlab – the digital humanities center at Umeå University – focusing on digital text analysis, media history and strategic communication.
• Väinö Yrjänäinen - doctoral student at the Department of Statistics, Uppsala University. Building on his educational background in data science and strong interest in natural language processing
• Robert Borges - linguist with a background in sociolinguistics, language description, and linguistic typology.

I feel you have now tested in 10 years to produce open data without creating new digital teams... maybe everyone inside the community are satisfied but from the outside it looks inefficient...

• Nationell dataverkstad is producing guidelines how to save from Excel to csv and explains csv if you have people in a project that understands csv you will never produce anything...

• I let chat GPT create a spec for "kommundiarier"

---

say how we can jump directly to all linked data and prove that it is possible across the board.

• ask chatGPT if you are unsure

Can you in simple language explain the importance of this list for a non skilled top manager

[bjornhagstrom]

That was not an answer to my question. Do you have examples where adopting a linked data approach has worked as a general principle when working with organisations that have a lot of legacy data and systems?

[Lonestarrk]

Additional fields in the api specification is not a bad thing even if the data isn't available now doesn't it won't ever be... more fields in the data contract doesn't beak anything for existing users and when the data is available there will be no confusion or need to change the specification. Some future implementor could have other possibilities than we do. If school ID has a use case I vote yes. Den tors 20 apr. 2023 15:51Magnus Sälgö ***@***.***> skrev:

…

- learn from the Humlab project <http://url> they have very skilled people Example of team members - Måns Magnusson <https://github.com/MansMeg> Uppsala Universitet assistant professor in Statistics at the Department of Statistics, Uppsala University <https://www.statistik.uu.se/?languageId=1> - Fredrik Noren <https://www.umu.se/personal/fredrik-noren/> working as a senior research assistant at Humlab – the digital humanities center at Umeå University – focusing on digital text analysis, media history and strategic communication. - Väinö Yrjänäinen <https://www.westac.se/medverkande/vaino-yrjanainen/> - doctoral student at the Department of Statistics, Uppsala University. Building on his educational background in data science and strong interest in natural language processing - Robert Borges <https://robertborges.net/cv/> - linguist with a background in sociolinguistics, language description, and linguistic typology. I feel you have now tested in 10 years to produce open data without creating new digital teams... maybe everyone inside the community are satisfied but from the outside it looks inefficient... - Nationell dataverkstad <https://www.vgregion.se/ov/dataverkstad/> is producing guidelines how to save from Excel to csv and explains csv if you have people in a project that understands csv you will never produce anything... - I let chat GPT create a spec for "kommundiarier" <https://gist.github.com/salgo60/8105808191095303aff3e3e8b6d8b445> ------------------------------ say how we can jump directly to all linked data and prove that it is possible across the board. - ask chatGPT if you are unsure <salgo60/DIGG-skuggbacklog#70 (comment)> [image: image] <salgo60/DIGG-skuggbacklog#70 (comment)> [image: image] <salgo60/DIGG-skuggbacklog#70 (comment)> Can you in simple language explain the importance of this list for a non skilled top manager [image: image] <https://user-images.githubusercontent.com/14206509/233386765-5ffb2318-6d0f-4647-b8f8-ef4207fee029.png> [image: image] <https://user-images.githubusercontent.com/14206509/233386820-798c586d-2087-402d-8c54-ce8bd93e2e55.png> [image: image] <https://user-images.githubusercontent.com/14206509/233386920-815aa28e-de82-4cc9-899f-54a1aa1467c5.png> — Reply to this email directly, view it on GitHub <#14 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/ACLM4LYF2OUSNHF34KSULPLXCE5NXANCNFSM5YTQXY7A> . You are receiving this because you are subscribed to this thread.Message ID: ***@***.***>

[salgo60]

@Lonestarrk
Its not rocket science to do reconciliation we have all schools and ids in #Wikidata

• "Skolenhetskod" wikidata P7894
  • list / map

Datastory created this data in Wikidata and did an application Skolkartan

• interactive tool

[bjornhagstrom]

@salgo60 That is still not an answer to my question.

[salgo60]

You have ELI see workshop - 2 min video

[bjornhagstrom]

I will take that as a "No".

[salgo60]

Yes, adopting a linked data approach has been successful for many organizations dealing with large amounts of legacy data and systems. Linked data is a method of publishing structured data on the web, allowing it to be easily interlinked and shared across different systems. This approach has been particularly useful for organizations looking to improve interoperability, increase data discoverability, and streamline data management. Here are a few examples:

British Library:
The British Library adopted a linked data approach to integrate and expose its diverse and extensive datasets. This allowed the organization to link and share data with other cultural institutions, enabling users to access and explore data in new and innovative ways. The British Library's use of linked data also facilitated collaboration with other libraries and cultural institutions, fostering a more interconnected information ecosystem.

BBC:
The British Broadcasting Corporation (BBC) implemented a linked data approach for its online content, integrating various data sources and systems to provide a more comprehensive and dynamic user experience. The BBC's linked data approach facilitated the linking of news articles, TV and radio programs, and other multimedia content, improving the discoverability and consumption of content across its digital platforms.

US Government:
Data.gov, the United States government's open data platform, has utilized linked data principles to improve interoperability and data accessibility. By using standardized formats and vocabularies, Data.gov enables agencies to share and integrate data more effectively, ultimately providing users with better access to government data and promoting transparency.

Europeana:
Europeana is a digital platform for cultural heritage, providing access to millions of digitized items from European libraries, archives, and museums. Europeana adopted a linked data approach to improve data integration and discoverability. By interlinking diverse datasets, Europeana enables users to explore related content across different institutions, thus enhancing the overall user experience.

Pharmaceutical Industry:
The pharmaceutical industry has also benefited from linked data approaches in managing complex data related to drug discovery, clinical trials, and patient data. By implementing linked data, organizations can more effectively integrate and analyze data from various sources, improving decision-making and driving innovation in drug development.

These examples illustrate how adopting a linked data approach can help organizations with legacy data and systems improve data integration, discoverability, and interoperability. This, in turn, enhances user experiences and promotes collaboration across different domains.

[bjornhagstrom]

For me that are examples where it has worked, not systemic changes to entire systems. There are no municipalities there either and municipalities are quite different from governmental agencies.

[salgo60]

Listen to the ELI workshop I feel Spain has implemented in a rather complex organisation video slides

• give examples what is so complex with municipalities? when they cant produce so easy data as outdoor gyms then they will have a problem with everything ....
• I think the main problem is that you never design an echosystem you think 290 municipalities should upload csv files and thats it
  • all those csv files are not generated from a common infrastructure instead it feels like a mess. One very bad example of this mess is that one law created 290 datasilos and I guess no skilled information architect is involved instead there are webmasters with no vision more than creating fancy webpages....
  • compare the wikipedia infrastructure link that supports > 200 languages, we can track what facts exists with sources and also track if more sources "contradict each other" or if some sources are not confirmed by "better sources.... and all this can be displayed in > 200 languages and have SPARQL endpoint

e.g. Arabic 2nd biggest language in Sweden with fallback Swedish

[bjornhagstrom]

Municipalities are responsible for a wast number of areas compared to a governmental agency that has a limited scope. Municipalities also have a lot less resources. That creates an entirely different environment and situation. We have trouble getting people to create css files and you are talking about sparql endpoints. You are not even on the right planet here. Well this is the last post from me here this time. We have had this conversation many times over now and as usual I agree with you that what you say is the best solution but I still live in the real world and have deal with that and work with that.

[salgo60]

Municipalities also have a lot less resources. That creates an entirely different environment and situation

I agree and then you need to build an infrastructure that is datadriven.... Wikidata has SPARQL, has structured data see how Humlab is using that data link.... BUT most of the people in Wikipedia are not rocket scientists, they dont know what a csv file is.... still it works.... --> a good infrastructure exist that supports versioning, > 200 languages, RDF, 5 star data, showing the data in > 200 languages see "#6 Nationell dataverkstaden bör skapa ett nationellt ekosystem hur data kopplas ihop 5 star" I feel dataverkstaden have problems adding structure on the data they themself produce how should they be able to create an infrastructure? We see the same lack of skills with DIGG... so the problem is in the steering committee of the dataverkstaden project that dont see the lack of skills and with DIGG the whole echosystem feels old school when a thing like SFS 1999:725 can go on for 26 years ...

You need an echosystem

• 10 year trying to produce Open data based on that text strings should be moved to a central portal by hand and then use DCAT-AP and send those text strings in Swedish to an international portal should be understood is wrong and that should have been seen day 1.... now I feel no one has a solution or skills.... see Fredrics data problem with data portal väst and out door gyms .... sending swedish text strings to EDP... its crazy

not systemic changes to entire systems

• the basic problem i feel @bjornhagstrom is that you need to understand linked data
  • if two people should speak then we use words to communicate
  • if two machines should communicate then we speak about semantic interoperability if you dont understand that and how to do it then you do as Fredric does with Dataporal Väst and sends to EDP a text string "Utegym Tidaholm" I call this antipattern #77-4 Strings not Things --> that system B needs to start guess... if you have been working with Open Data for more than 10 years and dont have basic understanding of semantic interoperability then we should start questioning the skills of the organisation... good videos on the subject
    * Europeana Linked open data What is it 5 min
    * TED Tim Berners-Lee: The next Web of open, linked data 17 min more technical

ELI has started to be implemented in Europe

When when we look at Sweden they started in 1986 to implement SFS 1989:725 and 26 years later it looks like they have given up... was it difficult no just producing the most important document 100 governments produce called "författningar" in the same way.... when I asked Regeringskansliet they dont know they are responsible for the project.... its a playground wasting tax payers money see todays tweet...

There are no municipalities there either and municipalities are quite different from governmental agencies

the reason is spelled L A C K of competence and I guess no one is interested to change the current system.... see if Niels will do something

You need skills

I asked chatGPT see also tweet

• tweet

Open data communities push using DCAT-AP to distribute the data. The anti-pattern I see is that "strings not Things" are distributed and controlled vocabularies are not used. Can you explain WHY cross posting strings in a multi language environment is a bad pattern and LOD is important.

What knowledge and visions does top management need to start delivering Linked data? Today when I asked the boss of the expert Swedish Government what a Knowledge graph is I get a feeling they dont understand what to deliver and what skills are needed

Do you have examples of such a roadmap

Can you give me a checklist to see if an organisation is Linked data ready?

Can you in simple language explain the importance of this list for a non skilled top manager

Lack of data quality I call metadata debt. Do you have a good structured approach how an organsation should add quality to its data

Cleaning your data I would say you need:

1. tracability of open issues
   versioning of your data that is possible for every one to see
2. a possibility for external consumers to add a change request
3. a possibility to subscribe on changes
4. API access
   Anything else?

• see this and more at "#70 DIGG bör vara tankeledande och kunna tekniker som kunskapsgrafer"

[salgo60]

More examples of why owl:sameas is needed and the cost of bad data is not ok 2023

@bjornhagstrom more examples why owl:sameas should be part of the spec and then we get 5 stardata see explaination below in Swedish why its good and just a text string doesnt make sense for machines to read..... I assume one use case is that schools which serve quality food and have a broader selection of dietary options might want to showcase this and would be willing to generate reliable data. In the long run, there should be interest from initiatives like "Väl godkänt för skolmaten" and politicians to monitor this

if you cant deliver 5 stardata you have the wrong team.... based on the experience with SFS 1999:175 / #5, it's evident that sometimes, even after 23 years, organizations can't deliver and remain non-learning entities so not changing the team is not ok

Now datastory produced www.berattarministeriet.se/skolkartan/

If you had a map of schools in Sweden with a "högt socioekonomiskt index" (high socioeconomic index) and also information about the food served in these schools, there are various interesting relations and patterns you can investigate: - TO GET THIS working you need owl:sameas si we understand its the same school

Nota bene: also the food served should be linked data to make it 5 star data and machine readable...

1. Quality and Variety of Food:

* Determine if schools with a high socio-economic index offer a wider variety of food options compared to those with a lower index.
 * Investigate if the quality of ingredients (e.g., organic, locally-sourced) correlates with the socio-economic status of the school's location.

2. Cultural Diversity of Meals:

* Examine if schools in areas with a higher socio-economic index offer a more diverse range of international cuisines, reflecting a potential multicultural environment.

3. Health and Nutrition:

* Investigate the nutritional value of the meals provided in schools with different socio-economic indexes. 
Do schools in higher index areas provide more nutritionally balanced meals?
* Analyze the prevalence of junk food or processed food options.

4. Dietary Options:

 * Examine the availability of vegetarian, vegan, gluten-free, and other dietary-specific options. This can indicate awareness and accommodation of diverse dietary needs and preferences in schools of varying 
 socio-economic backgrounds.

5. Economic Factors and Meal Costs:

 * Investigate if the cost of school meals differs based on the socio-economic index and whether more affluent areas subsidize meal costs more heavily.
 * Analyze the percentage of students who utilize free or reduced-price meal programs in relation to 
 the socio-economic index.

6. Local Agriculture and Sustainability:

 * Examine the connection between schools in high socio-economic areas and their support for local agriculture, f
 * arm-to-table initiatives, or other sustainability practices in their meal programs.

7. Educational Components:

 * Investigate if schools in areas with a high socio-economic index incorporate more food and nutrition education 
 into their curriculum.
 * Assess if there are school garden programs or cooking classes available and how they might 
 correlate with socio-economic status.

8. Feedback and Parental Involvement:

* Examine if there's a correlation between the socio-economic status and the level of parental involvement or feedback regarding school meals. Higher socio-economic status might result in more parental involvement, which could 
influence the quality and type of food served.

9. Food Allergies and Sensitivities:

* Analyze if schools in higher socio-economic areas are better equipped or more responsive to managing 
students' food allergies or sensitivities.

By investigating these relations, you can gain a deeper understanding of how socio-economic factors
might influence the quality, diversity, and healthfulness of food served in schools. This information can be
valuable for policymakers, educators, and parents alike.

[salgo60]

Do you have examples where adopting a linked data approach has worked as a general principle when working with organisations that have a lot of legacy data and systems?

chatGPT answer you