Analysing Translation of Wikidata Properties

John Samuel

CPE Lyon

(Based on presentation made at Wikidata Con 2019, Berlin)
Creative Commons License

Outline

Outline

  1. Introduction
  2. WDProp: Wikidata Property Analysis
  3. Towards a coarser analysis on property translation*
    • Data Collection
    • Analysis Results
  4. Conclusion and Future Works

* Work done with Thibaut Chamard

1. Introduction

1. Introduction

Wikidata Item: Labels, Descriptions and Aliases

Example Wikidata Item: Bogotá(Q2841)

1. Introduction

Wikidata Item: Labels and Properties

Example Wikidata Item: Avignon (Q6397)

1. Introduction

Wikidata Properties

Wikidata Properties

1. Introduction

Wikidata Property: Labels and descriptions

Example Wikidata Property: country (P17)

1. Introduction

Properties: Labels, description and aliases

Labels, description and aliases of Property:P31 in English

1. Introduction

Wikidata Property: from Proposal to Usage

Property proposal, creation and usage on Wikidata

1. Introduction

Wikidata Property: Translation

Property Creation on Wikidata

1. Introduction

Wikidata Property: Missing descriptions

Missing multilingual descriptions on Property P31

1. Introduction

Wikidata Properties: Missing descriptions

Missing multilingual descriptions on Property P31

1. Introduction

Wikidata Properties: Missing labels

Missing multilingual labels on Property P31

2. WDProp: Wikidata Property Analysis

WDProp

  1. WDProp:
    • Collaborative Multilingual Multi-domain Ontology development: is it possible to achieve a truly multilingual experience?
  2. Properties:
    • Not so much influenced by bots!
  3. Goals:
    • Understanding Wikidata property proposal, creation and translation
    • Available templates and their usage
    • Providing real-time statistics to (multilingual) contributors

2. WDProp: Wikidata Property Analysis

Towards 'Everything' about Wikidata properties

2. WDProp: Wikidata Property Analysis

Towards 'Everything' about Wikidata properties

Information on Wikidata Properties

2. WDProp: Wikidata Property Analysis

2. WDProp: Wikidata Property Analysis

Wikidata Properties

Wikidata properties

2. WDProp: Wikidata Property Analysis

Wikidata Properties

Wikidata properties: All included deleted ones

2. WDProp: Wikidata Property Analysis

Wikidata Properties: Multilingual Labels

Real time information on number of multilingual labels

2. WDProp: Wikidata Property Analysis

Wikidata Properties: Multilingual Descriptions

Real time information on number of multilingual descriptions

3. Analysis of Translation of Properties

WDProp: Translation path

Wikidata properties: Translation path of three Wikidata Properties

3. Analysis of Translation of Properties

WDProp: Translation path

Wikidata properties: Possible vandalism (removal of labels)

3. Analysis of Translation of Properties

WDProp: Translation path

Wikidata properties: Possible vandalism (removal of labels, descriptions and aliases)

3. Analysis of Translation of Properties

Wikidata Properties: Revision History

Revision History on Wikidata: Property P17

3. Analysis of Translation of Properties

Wikidata Properties: Comments from revision history

Attribute Add Update Remove
label wbsetlabel-add wbsetlabel-set wbsetlabel-remove
description wbsetdescription-add wbsetdescription-set wbsetdescription-remove
alias wbsetaliases-add wbsetaliases-add-remove wbsetaliases-remove

3. Analysis of Translation of Properties

Wikidata Properties: Comments from revision history

3. Analysis of Translation of Properties

Wikidata Properties: Data Collection

3. Analysis of Translation of Properties

Wikidata Properties: Data Collection

timestamp property language type
2013-10-26T16:45:20Z P999 en label
2013-10-26T16:45:20Z P999 en description
2013-10-26T18:16:05Z P999 de label
2013-10-27T02:55:15Z P999 pt label
2013-10-27T02:55:18Z P999 fr label
2013-10-28T12:42:28Z P999 pl label
2013-10-29T08:41:45Z P999 fr description

3. Analysis of Translation of Properties

Wikidata Properties: Analysis Results

Histogram on the distribution of number of languages for different properties

3. Analysis of Translation of Properties

Wikidata Properties: Analysis Results

Measure Value
count 6347
mean 21.444935
std 20.897675
min 1.000000
max 154.000000

3. Analysis of Translation of Properties

Analysis Results: Frequent itemsets

Count Language Combinations
6347 en
6338 ar
6306 nl
6263 uk
5928 fr
4823 ca
4019 de

3. Analysis of Translation of Properties

Analysis Results: Frequent itemsets

Count Language Combinations
6259 ar, uk
5923 fr, ar
5905 fr, uk
4819 ar, ca
4795 ca, uk
4530 fr, ca
4015 ar, de

3. Analysis of Translation of Properties

Analysis Results: Frequent itemsets

Count Language Combinations
6257 en, nl, ar, uk
5919 en, nl, fr, ar
5903 en, nl, fr, uk
5901 en, ar, fr, uk
5899 nl, fr, uk, ar
4818 en, nl, ar, ca
4794 en, nl, ca, uk

3. Analysis of Translation of Properties

Wikidata Properties: Analysis Results

Translation Path

4. Conclusion and Future Works

Wikipedia Infobox Properties

Existing English Wikipedia Infobox Properties of Programming Languages

4. Conclusion and Future Works

Wikipedia to Wikdiata

Importing structured data from Wikipedia Infoboxes to Wikidata

4. Conclusion and Future Works

Wikdiata to Wikipedia

Exporting data from Wikidata to multiple multilingual Wikipedia articles

Thank you

Questions?

References

  1. Kaffee, L. A., Piscopo, A., Vougiouklis, P., Simperl, E., Carr, L., & Pintscher, L. (2017, August). A glimpse into Babel: an analysis of multilinguality in Wikidata. In Proceedings of the 13th International Symposium on Open Collaboration (p. 14). ACM.
  2. Müller-Birn, C., Karran, B., Lehmann, J., & Luczak-Rösch, M. (2015, August). Peer-production system or collaborative ontology engineering effort: What is Wikidata?. In Proceedings of the 11th International Symposium on Open Collaboration (p. 20). ACM.
  3. Samuel, J. (2017) Collaborative Approach to Developing a Multilingual Ontology: A Case Study of Wikidata. In : Research Conference on Metadata and Semantics Research. Springer, Cham, 2017. p. 167-172.
  4. Samuel, J. (2018). Towards Understanding and Improving Multilingual Collaborative Ontology Development in Wikidata. In: WikiWorkshop 2018
  5. Samuel, J. (2018). Analyzing and Visualizing Translation Patterns of Wikidata Properties, CLEF 2018, Avignon, France, 10-14 September, 2018, Lecture Notes in Computer Science, vol 11018. Springer, Cham