What I hate with a vengeange is make work. DBpedia as a project retrieves information from all the Wikipedias, wrangles it into shape and publishes it. In one scenario they have unanimous support from one or more Wikipedias agreeing on the same fact and, they all may have their own references.
We should import such agreeable data without further ado. An additional manual step to import to Wikidata is not smart because manual operations introduce new errors. Arguably when there is no unanimous support manual intervention may improve the quality but given the quantity of the data involved, it means that a lot of data will not become available. THAT in and of itself has a negative impact on the quality of available data as well.
So what to do.. Harvest all the data that is of an acceptable quality, that is the data DBpedia accepts for its own purposes. Enable an interface where people verify the data where their project is challenged.
When we truly aim to engage people, we enable them to target the data they want to work on. I will happily work on scientists but do not expect me to work on "sucker stars". More than likely there will be people who care about soccer stars but not about "crazy professors".
Thanks,
GerardM
Showing posts with label DBpedia. Show all posts
Showing posts with label DBpedia. Show all posts
Thursday, September 26, 2019
Wednesday, September 25, 2019
With #DBpedia to the (data) cleaners
The people at DBpedia are data wranglers. What they do is make the most of the data provided to them by the Wikipedias, Wikidata and a generous sprinkling of other sources. They are data wranglers because they take what is given to them and make the data shine.
Obviously, it takes skill and resources to get the best result and obviously, some of the data gathered does not pass the smell test. The process the data wranglers use includes a verification stage as described in this paper. They have two choices for when data that should be the same is not; they either have a preference or they go with the consensus ie the result that shows most often.
For data wranglers this is a proper choice.. There is an other option for another day, these discrepancies are left for the cleaners.
With the process well described, the data openly advertised as available, the cleaners will come. First people akin to the wranglers, they have the skills to build the queries, the tools to slice and dice the data. When these tools are discovered, particularly by those who care about specific subsets, they will dive in and change things where applicable. They will seek the references, make the judgments necessary to improve what is there.
The DBpedia data wranglers are part of the Wikimedia movement and do more than build something on top of what the Wikis produced; DBpedia and the Wikimedia projects work together improving our movement's qualities. With the processing data generally available this will become even more effective.
Thanks,
GerardM
Obviously, it takes skill and resources to get the best result and obviously, some of the data gathered does not pass the smell test. The process the data wranglers use includes a verification stage as described in this paper. They have two choices for when data that should be the same is not; they either have a preference or they go with the consensus ie the result that shows most often.
For data wranglers this is a proper choice.. There is an other option for another day, these discrepancies are left for the cleaners.
With the process well described, the data openly advertised as available, the cleaners will come. First people akin to the wranglers, they have the skills to build the queries, the tools to slice and dice the data. When these tools are discovered, particularly by those who care about specific subsets, they will dive in and change things where applicable. They will seek the references, make the judgments necessary to improve what is there.
The DBpedia data wranglers are part of the Wikimedia movement and do more than build something on top of what the Wikis produced; DBpedia and the Wikimedia projects work together improving our movement's qualities. With the processing data generally available this will become even more effective.
Thanks,
GerardM
Sunday, September 22, 2019
Comparing datasets, bigger or better or it does not matter?
When Wikidata was created, it was created with a purpose. It replaced the Wikipedia based interwiki links, it did a better job and, it still does the best job at that. Since then the data has been expanded enormously, no longer can Wikidata be defined by its links to Wikipedia as it is now only a subset.
There are many ongoing efforts to extract information from the Wikipedias. The best organised project is DBpedia, it continuously improves it algorithms to get more and higher grade data and it republishes the data in a format that is both flexible and scalable. Information is also extracted from the Wikipedias by the Wikidata community. Plenty of tools like petscan and the awarder and plenty of people working on single items one at a time.
Statistically on the scale of a Wikidata, individual efforts make little or no impression but in the subsets the effects may be massive. It is for instance Siobhan working on New Zealand butterflies and other critters. Siobhan writes Wikipedia articles as well strengthening the ties that bind Wikidata to Wikipedia. Her efforts have been noticed and Wikidata is becoming increasingly relevant to and used by entomologists.
There are many data sets, because of its wiki links every Wikipedia is one as well. The notion that one is bigger or better does not really matter. It is all in the interoperability, it is all in the usability of the data. Wikipedia wiki links are highly functional and not interoperable at all. More and more Wikipedias accept that cooperation will get them better quality information for its readers. Once the biggest accept data as a resource to curate the shared data the act of comparing data sets is improved quality for all.
Thanks,
GerardM
There are many ongoing efforts to extract information from the Wikipedias. The best organised project is DBpedia, it continuously improves it algorithms to get more and higher grade data and it republishes the data in a format that is both flexible and scalable. Information is also extracted from the Wikipedias by the Wikidata community. Plenty of tools like petscan and the awarder and plenty of people working on single items one at a time.
Statistically on the scale of a Wikidata, individual efforts make little or no impression but in the subsets the effects may be massive. It is for instance Siobhan working on New Zealand butterflies and other critters. Siobhan writes Wikipedia articles as well strengthening the ties that bind Wikidata to Wikipedia. Her efforts have been noticed and Wikidata is becoming increasingly relevant to and used by entomologists.
There are many data sets, because of its wiki links every Wikipedia is one as well. The notion that one is bigger or better does not really matter. It is all in the interoperability, it is all in the usability of the data. Wikipedia wiki links are highly functional and not interoperable at all. More and more Wikipedias accept that cooperation will get them better quality information for its readers. Once the biggest accept data as a resource to curate the shared data the act of comparing data sets is improved quality for all.
Thanks,
GerardM
Monday, April 17, 2017
#Wikidata user story - #DBpedia, #death and #Federation
Federation between DBpedia and Wikidata became possible. As a consequence, the results of a query that runs on DBpedia can be linked to Wikidata.
Some time ago people at DBpedia created a wonderful query that shows differences between DBpedia and the Dutch and Greek Wikipedia. It received approval from the Dutch Wikipedia community.
With federation something much more interesting became possible; a federated query comparing Wikidata with one DBpedia at a time. When the query runs, current data from Wikidata and DBpedia is presented. When a Wikipedia associated with DBpedia changes, DBpedia may import the differences from a RSS-feed and consequently running the query again will show the latest differences.
Updating information about one particular type of statement like date of death, place of death or whatever, will always be based on the current differences.. Experiencing the results in this way is truly motivating. Federation is an instrument that can helps us improve the quality of either federated system.
Thanks,
GerardM
Some time ago people at DBpedia created a wonderful query that shows differences between DBpedia and the Dutch and Greek Wikipedia. It received approval from the Dutch Wikipedia community.
With federation something much more interesting became possible; a federated query comparing Wikidata with one DBpedia at a time. When the query runs, current data from Wikidata and DBpedia is presented. When a Wikipedia associated with DBpedia changes, DBpedia may import the differences from a RSS-feed and consequently running the query again will show the latest differences.
Updating information about one particular type of statement like date of death, place of death or whatever, will always be based on the current differences.. Experiencing the results in this way is truly motivating. Federation is an instrument that can helps us improve the quality of either federated system.
Thanks,
GerardM
Thursday, March 30, 2017
#Quality - #DBpedia and Kappa Alpha Psi
Kappa Alpha Psi is a fraternity of students and alumni. There is a Wikipedia article in English, a Commons category and a Wikidata item.
The information about Kappa Alpha Psi at Wikidata is based on the Wikipedia article. Information was added to the items for the members. This was done because in a related item it was found that the influence of fraternities and sororities is considerable. Concentrating for a moment on Kappa Alpha Psi has a secondary quality impact on what is of primary concern but when this is done for three such organisations, it quickly affects thousands of notable people.
When people find it of interest to add information about a membership to a Wikipedia article it has some impact. Having a category helps more to make the relevance of a Kappa Alphi Psi more visible. Adding this information to Wikidata is easy and it may show up in any language when membership information is part of a template.
DBpedia is a project similar to Wikidata. It harvests data from Wikipedias more consistently than Wikidata. Wikidata items are mapped to its internal items making it is possible to compare Wikidata with DBpedia.
When quality is an objective, when quality is to be improved effectively, the differences between DBpedia and Wikidata are an easy and one of the more obvious starting points. For some Wikipedias DBpedia updates are based on the RSS feed of the changes. So once a difference has been curated and changed in either Wikipedia or Wikidata, it results in an improved DBpedia entry and the desired improvement in quality. It does not need any math to understand this.
What we needed is a tool that uses these differences as input for a subset that is of interest to a Wikidata volunteer. That might be the Kappa Alpha Psi, The Black Lunch Table or whatever. Whatever can be defined with a query.
Thanks,
GerardM
The information about Kappa Alpha Psi at Wikidata is based on the Wikipedia article. Information was added to the items for the members. This was done because in a related item it was found that the influence of fraternities and sororities is considerable. Concentrating for a moment on Kappa Alpha Psi has a secondary quality impact on what is of primary concern but when this is done for three such organisations, it quickly affects thousands of notable people.
When people find it of interest to add information about a membership to a Wikipedia article it has some impact. Having a category helps more to make the relevance of a Kappa Alphi Psi more visible. Adding this information to Wikidata is easy and it may show up in any language when membership information is part of a template.
DBpedia is a project similar to Wikidata. It harvests data from Wikipedias more consistently than Wikidata. Wikidata items are mapped to its internal items making it is possible to compare Wikidata with DBpedia.
When quality is an objective, when quality is to be improved effectively, the differences between DBpedia and Wikidata are an easy and one of the more obvious starting points. For some Wikipedias DBpedia updates are based on the RSS feed of the changes. So once a difference has been curated and changed in either Wikipedia or Wikidata, it results in an improved DBpedia entry and the desired improvement in quality. It does not need any math to understand this.
What we needed is a tool that uses these differences as input for a subset that is of interest to a Wikidata volunteer. That might be the Kappa Alpha Psi, The Black Lunch Table or whatever. Whatever can be defined with a query.
Thanks,
GerardM
Thursday, March 03, 2016
#Wikidata - to share the sum of all data
The BBC featured an article about all the battles of the world in space and time. The cool thing is that the people from Nodegoat used data from Wikidata and DBpedia and consequently the article is not about Wikidata but about what you can do with its data.
Nodegoat enables people to bring data together and visualise it in a way so that people can comprehend it. The BBC article describes the "Geography of Violence map"; it includes data about all known battles ever that include place and time information. Most of the data is from DBpedia and some of it is from Wikidata.
There are two issues to consider;
Nodegoat enables people to bring data together and visualise it in a way so that people can comprehend it. The BBC article describes the "Geography of Violence map"; it includes data about all known battles ever that include place and time information. Most of the data is from DBpedia and some of it is from Wikidata.
There are two issues to consider;
- Nodegoat available for everyone
- curating data from DBpedia and Wikidata
When Nodegoat is available at Wikimedia Labs, it will enable people to come up with new ideas, presenting data that is otherwise obscure. They will find that much data is missing or wrong and at Wikidata they can improve the data as they see fit. Lab1100 is considering what license for Nodegoat and when we would make Nodegoat available at Labs, it would give them a firm push towards a free license.
DBpedia is a real strong project; it has gleaned a lot of data from Wikipedia and its quality is high. When there is an error in the data you can only change it in Wikipedia and wait for DBpedia to update. With Wikidata you can update the data itself. The data that is included was largely retrieved with similar routines used by DBpedia so it actually makes sense to collaborate with DBpedia and have them update Wikidata as well.
When the sum of all data is available to all and when many tools are available to leverage the data, more people will have incentives to improve the underlying data. Sharing the sum of all data and providing tools that make a difference is the best way to fulfill our mission.
Thanks,
GerardM
Saturday, June 06, 2015
#Wikipedia - its tyranny of sources
Let us keep it simple; a credible source is a source that that provides good information. There are several aspects to it.
- accurate
- unbiased
- timely
- complete
A lot of research went into "proving" that Wikipedia is a credible source. Typically this was done by subject matter experts that examined a subset of articles. Usually a lot of "sources" coincided with credibility.
The point was made repeatedly, Wikipedia is credible for its intended use.
Proving that Wikidata provides good information is different. It is incomplete; a lot of information is just not there. Twenty percent of its items provide no information. When sources are cited, it is Wikipedia that is referred to. Data like deaths in 2015 is no longer maintained. The pope, the bishop of Rome, has no longer the "Roman Catholic Church" as his faith.
Wikidata is immature and that is ok.
The approach to sources for Wikidata is tyrannical. It imposes sources for each statement and the suggestion is that this will make it credible. When you consider the sheer number of statements and the speed at which they increase it is impossible to see how that will work. It is well understood that having a source for a "fact" is one way to perpetuate a lie.
There is a better approach to proving Wikidata as a credible source. This is done in the same way as it was suggested that Wikipedia was credible; by comparing it to others like Brockhaus and Encyclopaedia Britannica. Much of the ground work is already there. We can compare to VIAF, to the GND. The point of this would not be that all our data is the same but that we care about quality and share the differences and research those.
While we make our data as accurate as we can, we can continue to work on making our data complete. Any credible source including Wikipedia and by inference DBpedia qualifies. They qualify because they help us complete our data and do this in a timely way.
This process of adding data is iterative. We add, we check and research the differences.
Every argument why we cannot do this or should wait denies Wikidata its credibility and worse the call to wait till the latest and greatest gadget comes online demotivates. The Wikipedia approach to sources is great for Wikipedia it just does not work for Wikidata.
Thanks,
GerardM
Thursday, June 04, 2015
#English #Wikipedia and those who died in #2015
It is wonderful news and I am grateful that the English Wikipedia decided to use Wikidata in stead of its own templates. It is wonderful, they have realised that the good people at Wikidata maintain that data..
I did it with pleasure, and I am thrilled to learn that the data is of high quality and up to date. I made it so because copying dates from English into Wikidata is easiest and therefore I always did them first the past two years.
The move of the English Wikipedia to use Wikidata data has been used as an excuse NOT to use the data harvested by DBpedia. That argument is wrong for three reasons.
I did it with pleasure, and I am thrilled to learn that the data is of high quality and up to date. I made it so because copying dates from English into Wikidata is easiest and therefore I always did them first the past two years.
The move of the English Wikipedia to use Wikidata data has been used as an excuse NOT to use the data harvested by DBpedia. That argument is wrong for three reasons.
- they update information in real time by harvesting from the RSS feed.
- they do this not only for the English Wikipedia
- they do not rely on a poor sod who adds info by hand
PLEASE, make use of the DBpedia data and please do consider the DBpedia people as part of our world. Let us learn from what they learned and make use of what they so gladly want to share with us. Share with us on our terms, share with us using our license.
Thanks,
GerardM
Wednesday, June 03, 2015
#Wikidata - Jurandyr Noronha died in 2015
Mr Noranha died on 2015-05-10. Mr Noranha was active in the movie industry so he is known in the IMDB database.
People die all the time but with the list functionality Jura1 created a page for "Brazilian people who died in 2015" dead Brazilians have an added significance. Mr Noranha qualifies..
At issue is that while Mr Noranha is known to Wikidata, it still takes a person to make a manual statement about his final day. I have done that for a year but I find that I am falling more and more behind. I do not even make a statement for Mr Noranha to be a Brazilian.
One thing that is not motivating is that I KNOW that this information is there for Wikidata to be had. The wonderful people at DBpedia really want to share this with Wikidata. Not much happens and as it ism my enthusiasm is wavering and my precious free time that is in demand for other things as well..
So what to do..
Thanks.
GerardM
People die all the time but with the list functionality Jura1 created a page for "Brazilian people who died in 2015" dead Brazilians have an added significance. Mr Noranha qualifies..
At issue is that while Mr Noranha is known to Wikidata, it still takes a person to make a manual statement about his final day. I have done that for a year but I find that I am falling more and more behind. I do not even make a statement for Mr Noranha to be a Brazilian.
One thing that is not motivating is that I KNOW that this information is there for Wikidata to be had. The wonderful people at DBpedia really want to share this with Wikidata. Not much happens and as it ism my enthusiasm is wavering and my precious free time that is in demand for other things as well..
So what to do..
Thanks.
GerardM
Saturday, June 07, 2014
#Dutch (National #Library, #DBpedia, #Wikimedia chapter) cooperation
So far #Wikidata did not have much in the sense of GLAM attention. When the Koninklijke bibliotheek, the Dutch (DBpedia and Wikimedia) chapter have anything to say about it, this will change.
Their aim is to bring the sum of all Dutch authors to Wikidata. You can imagine that a national library knows about these authors. You can imagine that the Dutch Wikipedia knows many of them and so does the Dutch DBpedia. The great thing of this cooperation is that DBpedia will identify those authors Wikipedia knows so that a clean list of new items can be added to completely include all these authors.
There have been several meetings about it, an RFC happened in Wikidata and importing all this data is the first objective. The mapping of attributes for authors to both DBpedia and Wikidata will be published and when this meets approval the import will happen.
Once it is all imported, DBpedia will be aware of all the changes to both Wikidata and the Dutch Wikipedia. It will keep its data up to date and when this finds approval, it can do this for Wikidata as well. There is no license issue as the Dutch DBpedia considers any updates "real time" and all updates happen under the license of the project involved in their own process.
When this project is completed, when all the details have been worked out, other collaborations between these four parties are possible. For instance, many authors received awards for their works or for a specific book and there may be more books..
Thanks,
GerardM
Their aim is to bring the sum of all Dutch authors to Wikidata. You can imagine that a national library knows about these authors. You can imagine that the Dutch Wikipedia knows many of them and so does the Dutch DBpedia. The great thing of this cooperation is that DBpedia will identify those authors Wikipedia knows so that a clean list of new items can be added to completely include all these authors.
There have been several meetings about it, an RFC happened in Wikidata and importing all this data is the first objective. The mapping of attributes for authors to both DBpedia and Wikidata will be published and when this meets approval the import will happen.
Once it is all imported, DBpedia will be aware of all the changes to both Wikidata and the Dutch Wikipedia. It will keep its data up to date and when this finds approval, it can do this for Wikidata as well. There is no license issue as the Dutch DBpedia considers any updates "real time" and all updates happen under the license of the project involved in their own process.
When this project is completed, when all the details have been worked out, other collaborations between these four parties are possible. For instance, many authors received awards for their works or for a specific book and there may be more books..
Thanks,
GerardM
Labels:
DBpedia,
Dutch chapter,
GLAM,
KB,
Netherlands,
Wikidata
Friday, March 28, 2014
#DBpedia & #Wikidata - Gerarda Rueter, a Dutch sculptor
At the Dutch #Wikimedia chapter, we had a talk with the Dutch DBpedia chapter. We want to cooperate and there are multiple things that we can do. DBpedia is excellent in harvesting information from Wikipedia. The Dutch DBpedia harvests from the Dutch Wikipedia.
For one of the first DBpedia projects, we want to set a repeatable example. It should be easy and the benefits need to be obvious. Mrs Gerarda Rueter typifies what we hope to achieve. We only knew that she was a sculptor and all the other information on the page was added by hand later. What we did was identify her as a human and as a female. We added the date and place of both her birth and death. We added her father. For him and Mrs Rueter's siblings, we added spouses when known.
When you present the data using Reasonator, relevant information becomes visible. What can be considered is that names are spelled the same in many languages. That however is probably for a next round when nationalities are included.
Thanks,
GerardM
Wednesday, March 19, 2014
#GSOC - #DBpedia and #Wikidata
The most interesting Google Summer of Code proposal I have seen for 2014 is this one.
4.7. Clean DBpedia datasets and import in WikidataAs far as I am aware they are still looking for a candidate to choose this project. This project is totally relevant and it is exactly the kind of repeatable process that Wikidata should have. By having this project done to DBpedia standards, it is ensured that the baseline will be stable and the results will be repeated regularly.
The student who will take this task will be responsible for two things:
Clean up the DBpedia errors based on the output of Databugger (http://databugger.aksw.org) or similar. With this the student will generate a more sparse but cleaner dump of DBpedia that will be of general use.
Communicate with the Wikidata community in order to coordinate the import of (parts of) the cleaned datasets and re-use the connections of DBpedia to fetch additional data for Wikidata import.
Mentors: Dimitris Kontokostas, Magnus Knuth (co-mentor)
Thanks,
GerardM
Thursday, February 06, 2014
#DBpediaAmsterdam - my slides
These are the slides as I used them at the DBpedia conference in Amsterdam. As you may infer from the furious pace of new functionality becoming available in the Reasonator, there was no time to add the hyperlinks.
The most important ones are:
- wikidata.org
- dbpedia.org
- reasonator.info
- page view statistics for Reasonator
- relevant Wikidata statistics
Thanks,
GerardM
Friday, January 31, 2014
#DBpediaAmsterdam - a confererence update
Speaking about #Wikidata at the #DBpedia conference in Amsterdam was a joy and a privilege. It was a joy because there was a real interest even a practical interest in what can we do to make a difference.
These are some of the points that they found most relevant
These are some of the points that they found most relevant
- Wikidata is used in templates in several Wikipedias
- for DBpedia, the specific properties in use in Wikidata are mostly a mapping issue
- having more properties in use than are on display in a Wikipedia is in Wikidata not an issue
- Wikidata is language agnostic while DBpedia has projects concentrating on specific languages
The process of how Wikidata is filled with information by bots was discussed at length. What bots do is harvest information from a Wikipedia and update the relevant items in Wikidata. This process is exactly the same for the update of DBpedia. The innovation that DBpedia can bring is by adding one extra functionality; reporting on the differences found. Obviously this will include new statements but identifying where Wikidata and Wikipedia have differences is really relevant when we work on improving the quality of the information we provide.
The relation of GLAM and Wikidata was another topic that received attention. In one presentation, Abraham Jacob van der Aa was mentioned. I added several statements for Mr van der Aa and showed how it shows in multiple languages in the Reasonator. It is not unlikely that as a result the Koninklijke Bibliotheek may provide us with information on Dutch authors in addition to what is available in DBpedia. The prospect of the Commons / Wikidata integration was also discussed; the existing links from Wikidata to Commons were mentioned but the possibility to link to the "Creator" and "Institution" templates were appreciated as having relevance to current practices.
The DBpedia conference had multiple simultaneous sessions so it should be obvious that much more was discussed. It is also my update of a conference that dealt with so much more and where a really vibrant community showed off what issues they are dealing with.
Thanks,
GerardM
The DBpedia conference had multiple simultaneous sessions so it should be obvious that much more was discussed. It is also my update of a conference that dealt with so much more and where a really vibrant community showed off what issues they are dealing with.
Thanks,
GerardM
Labels:
Commons,
DBpedia,
GLAM,
Netherlands,
Reasonator,
Wikidata,
Wikipedia
Wednesday, January 29, 2014
The #DBpedia experience
DBpedia has been a regular subject on this blog. It is the culmination of the work of many research projects and its data is based on what can be retrieved from Wikipedia. One of the results is that an impressive amount of resources know how to link to DBpedia.
Tomorrow there is a DBpedia conference in Amsterdam. One of the subjects is the future of DBpedia. As far as I am concerned, there is a bright future for the approach that made DBpedia a reality. A future that they will share with Wikidata. In the end, it is the Wikidata project could do with more cooperation.
Thanks,
GerardM
Sunday, December 29, 2013
#Wikidata - how many ontologies do we need?
When you check out items in Wikidata, chances are that they have links to one or more external resources. The picture to the right shows how DBpedia is connected to even more external resources. As you may know, DBpedia and Wikidata also have a very tight connection.
In a way, this can be compared to the many connections Wikidata has to the categories in the many Wikipedias, in Commons. All of them have structure of some kind. Information is inferred and acted upon. The category "United States Army Medal of Honor recipients" refers to people who received the Medal of honor.
When Wikidata is linked to external sources, these links have to be there for a reason. It can be to compare information or it can be to add missing information either way. Applications based on the ontology structure of the external sources can apply their logic on Wikidata.
Yes, there may be a purpose for yet another ontology structure in Wikidata. It may allow for implied information in tools native to Wikidata. But there is no point to all of them when we do not make use of them.
Thanks,
GerardM
In a way, this can be compared to the many connections Wikidata has to the categories in the many Wikipedias, in Commons. All of them have structure of some kind. Information is inferred and acted upon. The category "United States Army Medal of Honor recipients" refers to people who received the Medal of honor.
When Wikidata is linked to external sources, these links have to be there for a reason. It can be to compare information or it can be to add missing information either way. Applications based on the ontology structure of the external sources can apply their logic on Wikidata.
Yes, there may be a purpose for yet another ontology structure in Wikidata. It may allow for implied information in tools native to Wikidata. But there is no point to all of them when we do not make use of them.
Thanks,
GerardM
Monday, September 02, 2013
Updates
The president of Latvia
#DBpedia indicates the current political party for Mr Andris Bērziņš, separately his previous party affiliations are indicated.
Wikidata uses qualifiers; it is possible to indicate start and end dates for the different affiliations. This is superior however, including all the existing information from DBpedia in Wikidata will boost the usefulness a lot. Sorting out all the qualifications is something we cannot do with current software anyway.
The need for a mass merge in Wikidata
Many of the newly created records have been identified as being duplicates and have already been merged. The example I used in my blogpost of the genus Hersilia has been given the label of "long-spinnered bark spider". A genus is a group and consequently it should be a plural. Because of renames like this it has become more difficult to identify candidates for merges.
Thanks,
GerardM
#DBpedia indicates the current political party for Mr Andris Bērziņš, separately his previous party affiliations are indicated.
Wikidata uses qualifiers; it is possible to indicate start and end dates for the different affiliations. This is superior however, including all the existing information from DBpedia in Wikidata will boost the usefulness a lot. Sorting out all the qualifications is something we cannot do with current software anyway.
The need for a mass merge in Wikidata
Many of the newly created records have been identified as being duplicates and have already been merged. The example I used in my blogpost of the genus Hersilia has been given the label of "long-spinnered bark spider". A genus is a group and consequently it should be a plural. Because of renames like this it has become more difficult to identify candidates for merges.
Thanks,
GerardM
The President of #Latvia
#Wikidata needs more data to gain relevancy. At issue is where to get the data from. Some argue that it is to be preferred to retrieve data from the many #Wikipedias. This would be acceptable when the technology exists to retrieve the data.
The current pywikipedia functionality breaks easily and it does not take in all the available data. Consider Mr Andris Bērziņš who is the current president of Latvia. As you can read in the infobox, he was a member of the Communist Party before 1990. The bot will only identify him as a member of the Communist party of the Soviet Union. It does not add the qualifier (Before 1990). Mr Bērziņš is currently not affiliated with any political party.
The best way to present this information in Wikipedia is probably to have the most relevant information at the top. That would have helped the bot. For the bot it would be good if it accepted all the information until the next label. When the information that cannot be parsed goes into an "error file", the qualifiers can be added later by hand.
As it is not possible to reliably retrieve data from the Wikipedias, the arguments against using data from DBpedia lose their relevancy. This data is available for use warts and all and is a big improvement over the current lack of data and the inability to get the data out of the Wikipedias.
Thanks,
GerardM
The current pywikipedia functionality breaks easily and it does not take in all the available data. Consider Mr Andris Bērziņš who is the current president of Latvia. As you can read in the infobox, he was a member of the Communist Party before 1990. The bot will only identify him as a member of the Communist party of the Soviet Union. It does not add the qualifier (Before 1990). Mr Bērziņš is currently not affiliated with any political party.
The best way to present this information in Wikipedia is probably to have the most relevant information at the top. That would have helped the bot. For the bot it would be good if it accepted all the information until the next label. When the information that cannot be parsed goes into an "error file", the qualifiers can be added later by hand.
As it is not possible to reliably retrieve data from the Wikipedias, the arguments against using data from DBpedia lose their relevancy. This data is available for use warts and all and is a big improvement over the current lack of data and the inability to get the data out of the Wikipedias.
Thanks,
GerardM
Saturday, August 31, 2013
Mapping from #DBpedia
Mapping DBpedia properties to Wikidata may be difficult because a Vice President of the Unites States is defined by the same property as a Vice President of ....
"Obviously" you may deduce information from the context. Obvious when you have the technology.
Thanks,
GerardM
Tuesday, August 27, 2013
The tragedy in a #DBpedia announcement
There was an announcement of achieved goals in the Google Summer of Code for a DBpedia project.
The improved data quality is not intended to improve the quality of the data in the Wikipedia infoboxes. This is really sad because as the cooperation with the Deutsche National Bibliothek and the German Wikipedia proves, improvement is beneficial to both parties.
My understanding of why the Wikipedia infoboxes are out of scope is that there has to be a willingness to accept outside 'interference" and we have not learned to cooperate, not even among Wikipedia communities. Wikidata however is the new community and, it has a much higher standard to comply with. The data has to be superior to the data in any Wikipedia because it aims for its data to be used on any Wikipedia.
The best thing to do is side step this issue and make sure that all the questionable data in Wikidata is flagged so that we can start to find out what is best after all.
Thanks,
GerardM
Wikidata integration inside DBpedia we are happy to announce that an initial RDF DBpedia Dumps for Wikidata Data is now available.Enough reason to read again what this integration is about. The sad thing is in the Additional goals section; "Data quality assurance and tests to improve data quality for academia and enterprises (not for wikipedia infoboxes) would be nice".
The improved data quality is not intended to improve the quality of the data in the Wikipedia infoboxes. This is really sad because as the cooperation with the Deutsche National Bibliothek and the German Wikipedia proves, improvement is beneficial to both parties.
My understanding of why the Wikipedia infoboxes are out of scope is that there has to be a willingness to accept outside 'interference" and we have not learned to cooperate, not even among Wikipedia communities. Wikidata however is the new community and, it has a much higher standard to comply with. The data has to be superior to the data in any Wikipedia because it aims for its data to be used on any Wikipedia.
The best thing to do is side step this issue and make sure that all the questionable data in Wikidata is flagged so that we can start to find out what is best after all.
Thanks,
GerardM
Subscribe to:
Posts (Atom)















