View source for Web data integration

'''Web data integration''' (WDI) is the process of aggregating and managing data from different websites into a single, homogeneous workflow.  This process includes data access, transformation, mapping, quality assurance and fusion of data. Data that is sourced and structured from websites is referred to as "web data". WDI is an extension and specialization of [[Data integration|data integration]] that views the web as a collection of heterogeneous databases.
Data integration techniques in the context of the web, forms the foundation for businesses taking advantage of data available on the ever-increasing number of publicly-accessible websites.<ref>{{Cite web|url=https://www.uni-mannheim.de/dws/teaching/course-details/courses-for-master-candidates/ie-670-web-data-integration/|title=IE 670 Web Data Integration|date=2019-01-24|website=www.uni-mannheim.de|language=en|access-date=2019-02-11}}</ref> Corporate spending on this area amounted to about USD 2.5bn in 2017, and it is expected that by 2020 the market will reach almost USD 7bn.<ref>{{Cite web|url=http://www.opimas.com/research/355/detail/|title=Opimas: The Web Data Extraction Market|last=|first=|date=|website=Opimas: We begin with an understanding|language=en|archive-url=|archive-date=|access-date=2019-02-12}}</ref>

== Sources ==
Web data integration extends and specializes data integration to see the web as a collection of views of databases accessible over the web protocols, including, but not limited to:<ref>{{Cite web|url=https://www.webdataintegration.io/wdi/1.0/index.html|title=Introduction :: Web Data Integration|website=www.webdataintegration.io|access-date=2019-02-14}}</ref>

* Open data catalogs
* Government data catalogs
* Web applications and sites
** UI (web scraping)
** API
* The semantic web (SPARQL)
* HTML embedded structured data
* HTML data tables
* Spreadsheets
* PDFs
* Online encyclopedias

== Data access and transformation ==
WDI has technical challenges different from data integration due to the data access and transformation required for the web data sources being often [[Unstructured data|unstructured]] or [[Semi-structured data|semi-structured]] data without a standard query mechanism.

== Data quality ==
Understanding the quality and veracity of data is even more important in WDI than in data integration, as the data is generally less implicitly trusted and of lower quality than that which is collected from a trusted source. There are attempts to try to automate a trust rating for web data.<ref>{{Cite book|last1=Giménez-García|first1=José M.|last2=Thakkar|first2=Harsh|last3=Zimmermann|first3=Antoine|date=2016|editor-last=Sack|editor-first=Harald|editor2-last=Rizzo|editor2-first=Giuseppe|editor3-last=Steinmetz|editor3-first=Nadine|editor4-last=Mladenić|editor4-first=Dunja|editor5-last=Auer|editor5-first=Sören|editor6-last=Lange|editor6-first=Christoph|chapter=Assessing Trust with PageRank in the Web of Data|title=The Semantic Web|volume=9989|series=Lecture Notes in Computer Science|language=en|publisher=Springer International Publishing|pages=293–307|doi=10.1007/978-3-319-47602-5_45|isbn=9783319476025|chapter-url=https://hal-emse.ccsd.cnrs.fr/emse-01310508/file/JMGFG2016.pdf }}</ref>

Data quality in data integration can generally happen after data access and transformation, but in WDI quality may need to be monitored as data is collected, due to both the time and the cost of re-collecting the data.<ref>[https://metric.vn/ The way to integrate data from E-commerce website platform]</ref>

== Applications ==
WDI has application in many fields, including bioinformatics,<ref>{{Cite web|url=https://dbs.uni-leipzig.de/en/projekte/DATAINT/index.html|title=Web Data Integration|last=|first=|date=|website=Database Group Leipzig|archive-url=|archive-date=|access-date=}}</ref> search engines,<ref>{{Cite web|url=http://www.datascienceassn.org/content/web-scale-data-integration-you-can-only-afford-pay-you-go|title=Web-scale Data Integration - You Can Only Afford to Pay as You Go|website=www.datascienceassn.org|access-date=2019-02-12}}</ref> price comparison,<ref>{{Cite journal|title=Enabling global price comparison through semantic integration of web data|last1=Siegel|first1=Michael D.|last2=Madnick|first2=Stuart E.|date=2008|journal=International Journal of Electronic Business|language=en|last3=Zhu|first3=Hongwei|volume=6|issue=4|page=319|doi=10.1504/IJEB.2008.020672|hdl=1721.1/40084|s2cid=7995576|hdl-access=free}}</ref> and forensic search<ref>{{Cite web|url=https://www.consultancy.uk/news/2840/pwc-buys-kusiri-londonbased-fraud-detection-startup|title=PwC buys Kusiri, London-based fraud detection start-up|website=www.consultancy.uk|language=en|access-date=2019-02-12|date=2015-10-30}}</ref> data analysis, business intelligence, ecommerce,<ref>{{Cite book|last1=Osial|first1=P.|last2=Kauranen|first2=K.|last3=Ahmed|first3=E.|title=2017 IEEE 30th Canadian Conference on Electrical and Computer Engineering (CCECE) |chapter=Smartphone recommendation system using web data integration techniques |date=April 2017|chapter-url=https://ieeexplore.ieee.org/document/7946845/;jsessionid=GoG6IfE94h0FKAxRzD-ndyd2QUyOECVI7Pj9xc2zBt0s7mIuttf_!878820695|pages=1–5|doi=10.1109/CCECE.2017.7946845|isbn=978-1-5090-5538-8|s2cid=12474147 }}</ref> healthcare, pharmaceutical<ref>{{Cite web|title=How Data Integration is Revamping Healthcare and Pharma|url=https://dataintegrationinfo.com/data-integration-in-healthcare-pharma/|date=2020-04-27|website=Data Integration Info|language=en-US|access-date=2020-05-04}}</ref> and product development.

Most price comparison engines and recommendation systems use user generated data to create recommendations for their users. Similarly, healthcare systems use results of competitions conducted on websites like Kaggle<ref>{{Cite web|title=Kaggle: Your Machine Learning and Data Science Community|url=https://www.kaggle.com/|website=www.kaggle.com|language=en|access-date=2020-05-04}}</ref> to see the accuracy of data and to create user-focused products. In fact, IBM estimates that poor quality WDI is costing companies over $3 trillion<ref>{{Cite web|title=Web Data Integration: Revolutionizing the Way You Work with Web Data|url=https://www.import.io/post/web-data-integration-revolutionizing-the-way-you-work-with-web-data/|last=Import.io|website=www.import.io|language=en-US|access-date=2020-05-04}}</ref> in revenue each year.

==References==
{{Reflist}}

[[Category:Data management]]

{{Sourceattribution|Web data integration}}