Vmesnik za dostop do portala odprtih podatkov Slovenije

MARIĆ, SAŠO (author), Curk, Tomaž (mentor)

Portal Odprti Podatki Slovenije (OPSI) je vzpostavilo Ministrstvo za javno upravo s ciljem zagotoviti celostni popis podatkovnih zbirk, ki jih vodijo organi javnega sektorja, ter omogočiti objavo zbirk v obliki odprtih podatkov. Vsi podatki na portalu so podatki, evidence in zbirke, ki nastajajo pri delu organov javnega sektorja in so prosto dostopni. V diplomski nalogi smo obdelali dostopne podatke in jih pripravili v obliki za uvoz v program za strojno učenje in vizualizacijo podatkov Orange. Prikazali smo tudi statistiko metapodatkov na portalu OPSI. Vmesnik uspešno pretvori 788 datotek od 19565. Največ virov datotek in sicer 375 izhaja iz področja Vlada in javni sektor. V datotekah s preglednicami je 1038 stolpcev oz. spremenljivk, ki jih delimo v štiri skupine: diskretne (464), nizi (257), zvezne (180) in časovne znamke (137). Večina datotek (13765 datotek), ki jih vmesnik ne more pretvoriti samodejno, je formata html. S pomočjo vmesnika lahko portalu OPSI sporočamo pripombe ali izboljšave glede podatkov in javljamo nepravilnosti na portalu in podatkih. Podatke, ki jih razviti vmesnik samodejno pretvori, lahko uporabimo v programskem orodju Orange in v njih odkrivamo zakonitosti in zanimive vzorce.

• Language: Slovenian
• Keywords: odprti podatki, obdelava podatkov, pridobivanje spletnih podatkov
• Work type: Bachelor thesis/paper
• Organization: FRI - Faculty of Computer and Information Science
• Year: 2018
• PID: 20.500.12556/RUL-99917
• Publication date in RUL: 22.02.2018

Secondary language

• Language: English
• Title: Interface for accessing the portal Open data of Slovenia

Abstract

The portal Odprti Podatki Slovenije (OPSI) was established by the Ministry of Public Administration with the aim to provide an integrated listing of databases managed by the bodies of the public sector and to allow easy publication of data collections in form of open data. All the data provided on the portal are records and collections, which are created by the public sector bodies and are freely accessible. We have preprocessed the freely accessible data and formatted it into the format for Orange data mining, which is a program for machine learning and data visualization. We provide statistics on the available data in the OPSI portal. The interface automatically transforms 788 files out of 19565. The largest source of files, namely 375, is the area of the Government and the public sector. The interface automatically converts data in 1038 columns total, which are grouped into four categories: discrete (464), string (257), continuous (180) and time (137). Most of the files (13765 files) that the interface cannot transform automatically are in the html format. With the interface we can provide comments or improvements regarding the published data and report any irregularities of the portal and data. The data provided by the interface can be mined for interesting patterns using the Orange data mining software.

• Keywords: open data, data processing, web scraping