Education In Trentino - Data Integration

Reusability has a vital role in the Data Integration process defined by the ITelos methodology. This project is about the inegration of multiple hetegenious data sources following the standard Data Integration tools and technologies in order to produce useful expolitable knowledge graph maximizing the reuseability of the resources used. In the followingsections, you can find the details of the resources used including the breif description of each of them and links to the resources.

Scope of the project was a “A service that will help parent and student to find schools, including details about the school and courses offered, in the region of Trentino based on city, commune, school type, course duration and teaching activities schedules.

Datasets

Collection of data that supports the purpose of the project was one of the challenging part of this work. OPENdata Trentino was use to collect the data sets related to the education data in Trentino. It is a web portal that contains huge amount of public datasets owned by Autonomous Province of Trento.

Link of Autonomous Province of Trento

From the above mentioned portal, following datasets downloaded: • First dataset that was downloaded is “Trentino educational institutions” . The dataset is in XML format. It contains List of educational institutions in Trentino, with information such as, type of institution, site, training offer, contact information, address, school type, number and list of course offered.

Link of Trentino educational institutions

Second data set downloaded from the website was “Study courses of the Trentino schools” that contains list of study courses of Trentino schools with description and status of activity.The dataset is in XML format. This dataset has extra information, like starting year and duration, about the study courses being taught at the school.

Link of Study courses of the Trentino schools

List of all the communes in the province of Trento was manually downloaded from a website and cleaned.

Link of Trentino Comune Data

Knowledge Resource + Schemas

Follwing were the reference teleologies initially collected to satisfy the purpose of the project along the integration process :

• https://schema.org/EducationalOrganization
• https://schema.org/School
• https://schema.org/ElementarySchool
• https://schema.org/HighSchool
• https://schema.org/MiddleSchool
• https://schema.org/Preschool
• https://schema.org/City
• https://schema.org/Person
• https://schema.org/Course

ER Model

Based on the knowledge and data resources collected,follwing was the ER modelthat was created on the knowledge level which was later modified and also mapped to the refined datasets using data mapping tools.

Link to ER Model

ETG Model

From the model, ETG were created that define the conceptual relation of the entitites that are involved in the scope of the project. ETG in form of an OWL file that was created using the Protege tool which provides great support in this. Link of the ETG Folder

Knowledge Graph

The goal of the data integration project was to construct a solution that can answer CQs (competency questions). Unfortunately due to some unforeseen errors, we were not able to explored the final graph in the GrapghDb or any other similar tool for graph exploration. Invalid Base IRI error was encountered while importing the files in the GraphDb. We tried to fix the issue by redoing the mapping in the karmalinker and again generated the RDF graph files but this did not fix the issue. Despite doing the necessary research, we could not track down the issue. As result of this we were not able to properly explore the finally produced Knowledge Graph. We were able to import only part of graphs in the GraphDB that are also shown in the figures below.

Sample Graph Image 1 Sample Graph Image 2

Associated properties to each of the concepts can also be seen on the detailed widget on the right side. Despite this major setback, the analysis of the potential usage of the KG can be done on the conceptual bases following the modeling/entity matching done in the karmalinker. We tried to answer CQs following the incoming/outgoing links and in most of the cases we were able to find the desired data. For example, if someone wants to search all schools in Trento, we can very well start from the comune data-set looking for Trento comune and then following its conceptual link in the schools dataset to find out link of all school that are located in Trento comune. All of this reasoning was done manually without use of any tool due to the reason mentioned above. A possible solution of this problem could be by making sure that there is no entity that is un-mapped or improperly mapped in the karmalinker and then generate the RDF graphs. We could not do this as per time limitation and limited availability of the tools involved these phases of the project.

Final outout of the graphs can be found in the folder following the link given below. RDF Graph Outputs Folder

GITHUB Repository

All of the mentioned work can be found in the below github respository (master branch). Github Reposity Link - KDI Education In Trentino