Blog - www.ownyourdata.eu

Diabetes Data Processing

28. July 2020/0 Comments/in OwnYourData, SemCon /by Christoph Fabianek

In the NGI funded MyPCH project OwnYourData developed several technologies for secure and traceable data exchange of diabetes data: Digital Watermarking, Semantic Annotation, and Data Traceability. In addition, we also participated in a MyData Health initiative to write a feature article for the European Medical Writers Association (EMWA) and we are proud to announce that the article written by 14 individuals from 9 countries around the 6 MyData principles is already published and public available via the EMWA journal website – see section “Data Interoperability” and “Establishing trust between stakeholders for health data use” for example use cases of Semantic Containers.

In this blog post we cover the successful integration of diabetes data into the OwnYourData Data Vault. In this data flow, persons with diabetes (Pwd) can not only transfer their data to a Personal Data Store but also perform SPARQL queries to combine their diabetes data with public information – more information and examples are available here.

A special feature in the OwnYourData Data Vault is the Personal Knowledge Graph shown on the left side of the main screen. It compiles available data from the respective user and presents the information in a clear form. The screenshot below shows information about recent GPS Data, blood sugar levels over the course of a few days, as well as record numbers overall.

Beyond information in the Personal Knowledge Graph plugins allow further data exploration. In the course of the MyPCH project however, we decided to use existing tools like R or Jupyter Notebooks to provide more sophisticated visualization and analysis mechanisms. The R-Notebook available on Github is an example how to retrieve and decrypt information in the Data Vault and compile a report.

If you have any questions using Diabetes data with Semantic Containers or within the OwnYourData Data Vault don’t hesitate to contact us as support@ownyourdata.eu.

NGI Funding for DECTS

18. May 2020/0 Comments/in OwnYourData, SemCon /by Christoph Fabianek

The project DECTS (Deaf Emergency Chat and Training System) aims to provide deaf emergency calls and a training environment in several languages. With the help of a chatbot, deaf people can learn how to use the app and at the same time generate test data for training the control center personnel. The users can determine whether the entries are used as test data and there is documentation about origin, GDPR-compliant provision, and use of the data. The consent to the use of the data can also be changed and revoked later.

The teams from OwnYourData and DEC112 work together to implement this. In previous projects, an infrastructure was already set up in Austria for deaf emergency calls and the challenge now lies in international operations – for example when an Austrian tourist is on vacation in Copenhagen: in this case an emergency call is made via the DEC112 App registered in Austria and conveyed to the control center in Copenhagen.

A chatbot is developed for the training environment that simulates a control center. In a test chat, structured information is queried and further questions arise when using certain keywords. In cooperation with emergency call centers, typical conversations were analyzed and so-called decision trees were created, which the chatbot automatically processes.

If a user consents to the further use of the chat protocol, this consent can be managed in the OwnYourData Data Vault. There the consent of the transfer of the data is documented and it is possible to query when the data was accessed. In particular, however, access can also be restricted or subsequently prohibited. Semantic containers are used as the technology platform, which ensure data access is transparent and traceable.

Finally, personal data (emergency contacts, medical data and other information) can also be stored in the OwnYourData Data Vault, which is automatically provided to the control center in the event of an emergency chat. This personal data is referenced using a DID (Decentralized ID) and the data itself is stored encrypted. The Shamir’s Secret Sharing scheme ensures that the data can only be read by the user and the control center, but cannot be accessed by OwnYourData or DEC112.

The system architecture for the project is shown in the graphic below, together with the data flows between the individual components. All parts are now at least available as prototypes and the first end-to-end tests were carried out in May.

Semantic Technologies for Semantic Container

20. March 2020/0 Comments/in OwnYourData /by Christoph Fabianek

The vast amount of data being produced everyday requires semantics to provide meanings to the data. The recent advances of Semantic Web technologies provides us with standard data formats (i.e., Resource Description Framework – RDF), vocabularies (e.g., PROV-O for provenance) and tools (e.g., SPARQL for querying data) to add semantics to data and structuring the metadata. This use of semantics would in turn allows advances data analysis and processing to the data and its metadata.

Semantics for Diabetes Data

The MyPCH project aims to develop methods and tools to allow diabetes patients to share their data using Semantic Container. As part of the process, we provide a mean to transform the raw JSON data of diabetes patients’ observation data into RDF. The transformation is conducted in two steps: (i) Ontology definition, and (ii) Raw data transformation

We define a MyPCH ontology for the patient’s observation data based on the HL7-FHIR ontology for patient observation. The ontology is designed to maintain a high-degree of compliance with the original FHIR model while reducing the needs for data duplications required in the original FHIR-RDF model. An excerpt of the adapted ontology is shown in the figure below.

Figure: the FHIR-based ontology for the MyPCH project (click to enlarge)

As the next step, we utilize the RML mapping and CaRML engine to transform the raw JSON data into its RDF. The current version of RML mapping is available in Semantic Container. The resulted RDF representation of patients’ data would allow further analysis of the data using SPARQL queries (see also the tutorial for using these technologies in the semcon/sc-sparql container and the following dataflow with concrete diabetes examples). Furthermore, it would allow inferring additional knowledge based on the underlying knowledge with logical reasoning as well as integration of additional background knowledge (e.g., information about devices used for diabetes monitoring) and external knowledge, e.g., WikiData and DBPedia.

Extending SPECIAL Vocabularies for MyPCH user consent

Semantic Containers utilizes RDF-based vocabularies to represent its metadata. It adopts the standard Provenance Ontology (W3C PROV-O) to represent its provenance information and SPECIAL vocabularies to represent user consent, in addition to a custom RDF vocabularies to represent container’s metadata. We have extended the user consent vocabularies (SPECIAL vocabularies) with a set of additional classes specific for Semantic Container use cases using a specific namespace scp: <http://w3id.org/semcon/ns/policy#>. We utilize the vocabulary to automatically check the compliance between user consent of their data and the possible usage of the data by a data processor.

In the MyPCH project, we extend it further with three additional classes for diabetes patients’ data. The classes are assigned as specific types of SPECIAL health data category (svd:Health). These classes are for now scp:Diabetes (i.e., the generic diabetes patients’ data), scp: DiabetesSensor (i.e., insulin data observation from sensors), and scp:InsulinPump (i.e., insulin data observation provided by insulin pumps).

Personium Collaboration

18. December 2019/0 Comments/in OwnYourData /by Christoph Fabianek

Personium and OwnYourData are proud to announce a partnership for collaborating in Personal Data Store interoperability.

During the MyData 2019 conference (September 2019) in Helsinki, several PDS (Personal Data Store) innovators had come together and discussed openly regarding the interoperability between different PDSs. After the conference, Christoph Fabianek (OwnYourData) and Salman Farmanfarmaian (Freezr) contributed a first draft of the implementation document: CEPS – Common Endpoint for Personal data Stores.

The goal of the initiative is to allow users of different PDSs (right now: Freezr, Personium, Datafund, Data Vault) to use the same app (i.e., data processing capabilities) and switch between data stores. In a first step a simple app (Tally Zoo: maintain a tally chart for frequent tasks and collect your everyday data) was implemented that can connect natively to the individual PDSs. Subsequently, common interfaces for authorization and common operations like read/write should be harmonized across participating organizations to decouple data store development and app development.

Feel free to contact Dixon or Christoph if you have any questions or want to join the effort. We expect to showcase results in the upcoming MyData 2020 events.

Digital Watermarking

29. November 2019/0 Comments/in SemCon /by Christoph Fabianek

A digital watermark is a kind of marker covertly embedded in data and is also sometimes referred to as “the practice of imperceptibly altering a work to embed a message about that work”. For Semantic Container a digital watermark is a unique digital fingerprint that is applied to data provided by a Semantic Container, i.e., any data request results in a dataset with insignificant errors that uniquely identifies the recipient of the data set. In case such a dataset is leaked and appears in an unintended location, the person who originally requested and leaked the dataset can be identified. This blog post describes the design of the digital watermarking that will be implemented in the course of the currently ongoing MyPCH project.

Watermark embedding

To embed a watermark into a dataset the following two steps are performed:

Pre-processing: the available data is split up into fragments of a defined size, e.g., all measurements from a single day
Encoding: based on a secret parameter (or key) unique to the requesting party a sequence of errors with the same size as a data fragment is created and then applied to the original data, i.e., for numerical values this is just adding value and error

Watermark attacks

There are a number of possible attacks against digital watermarking:

Distortion Attack: There are different kind of distortions which may be applied to a dataset, e.g., rounding to the n-th digit. Rounding the values on the least significant digit preserves the data’s usability the most but may be detected more easily than rounding digits further up.
Deletion Attack: As with distortions attacks, different kind of deletions may be applied to a dataset to make the identification of the original recipient harder.
Collusion Attack: A collusion attack is performed by combination of n copies of the same dataset. For each measurement the mean of all n copies is calculated to create a new dataset.

Watermark Detection

To detect a watermark in a suspicious dataset the following two steps are performed and require the original data to be available:

Detection: Through similarity search the suspicious dataset (already fragmented) is matched against original data fragments and in case of a match the difference between suspicious dataset and original dataset is the (possibly noisy) unique error
Mapping: The extracted error is compared through similarity search with the original error based on the secret parameter (or key). In case of a match the original recipient of the data is identified.

The above process including various test cases for attacks will be implemented in the next weeks and will soon be available in the Semantic Container base package. Feel free to reach out to us with any questions or comments!

Milestone 3/3 – Summer

23. June 2019/0 Comments/in OwnYourData /by Christoph Fabianek

Semantic Containers for Data Mobility has reached all project goals by June 2019 and is now finalized with the final report. The three prototypes have developed well and Semantic Containers will continue to be used as infrastructure, especially at ZAMG.

As in the previous posts we summarize here our accomplished goals for this last milestone:

Seismic Data Use Case – basic demonstration of Semantic Containers
- source code: https://github.com/sem-con/sc-seismic
- available online: https://vownyourdata.zamg.ac.at:9500/api/data?duration=7
TAWES Use Case – Semantic Container with billing workflow
- source code: https://github.com/sem-con/sc-tawes
- available online (requires payment!): https://vownyourdata.zamg.ac.at:9600/api/data
Sentinel Use Case – process pipeline with Semantic Containers
- source code: https://github.com/sem-con/sc-sentinel
- available online:
  - download: https://vownyourdata.zamg.ac.at:9700/api/data
  - atmospheric correction: https://vownyourdata.zamg.ac.at:9701/api/data
  - clipping: https://vownyourdata.zamg.ac.at:9702/api/data
and also in the last 3 months there were again 18 project meetings.

Dates, dates, appointments!

SemCon was presented at the PyDays 2019 on May 3rd and we submitted a scientific publication at the Semantics 2019 in Karlsruhe.
Our final project presentation took place on 17 June 2019 at the MyData Meetup at the WU Vienna (slides are available here). We were happy about a lively participation and lively discussions on the topic “Data as Propery or Human right”. That’s what MeetUps should be like!
Another presentation is planned on September 27, where Christoph Fabianek will give a talk at the MyData 2019 with Diabetes.services from Denmark. We will continue to work on the topic of Semantic Containers even beyond the end of the funded project period!

In summary, we can look back on 9 successful project months. SemCon was presented around 10 times at events, we had about 60 internal project meetings, we were able to meet our milestones in time and there are 3 prototypes that show how the exchange with data can work simply and safely: by packaging data with usage rights and provenance information together with the necessary processing logic to enable reliable and easy management.

Latest News