Integration of Parking Sensory Data into UrbanSense

[thcarsten]

1. Use Case Title:

• Integration of Parking Sensory Data into UrbanSense

2. User Story/Scenario:

• As a data engineer, I integrate sensor data from different service providers, locations and modalities to the online platform 'UrbanSense', so that cities can analyse this data for their purposes. As a concrete example, for better planning of parking availibility throughout the city, parking and traffic data may be integrated, and visualized on a map.

3. Problem Statement:

• Currently, discovery of sensor data is a manual process and has to be repeated for each new city that is added to UrbanSense. For each use case (e.g. parking planning), relevant datasets need to be discovered manually. As such, even if generic solutions for using this data are available (e.g. a data visualization pipeline that works in a plug-and-play fashion), datasets have to be manually added to the existing data flows.

• Manual discovery of sensor data is time-consuming and often depends on business knowledge. As a data engineer without knowledge of the city infrastructure I may miss important sensor data sources available in the respective city, or need a person with business knowledge (e.g. a city employee) to assist in this data discovery.

4. Desired Outcome/Goal:

• One core goal of UrbanSense is to reduce costs of cities for building their digital infrastructure through sharing of resources between cities. Automation is another key element for reducing costs through improved efficiency. Discovery of sensor datasets can be largely automated if these datasets can be queried based on their content and relevance for the respective use case.

5. Data Catalog(s) Involved:

• Open Data Portaal Stad Gent: https://data.stad.gent/explore/

6. Data/Service Requirements:

• I want to be able to query an open data portal of a city like Gent so that I can discover any dataset that is related to traffic and parking within a range of postal codes. I want to be able to filter those datasets based on spatial and temporal criteria and vehicle-type. A nice-to-have would be that datasets results are sorted by relevancy for my particular use case (i.e. match with my provided prompt).

• For the Open Data Portaal Stad Gent, SHACL-shapes do not exist, because data is not provided in an rdf-format.

7. Discovery Criteria/Filters:

• Keywords: "parking", "occupancy", "availibility", "traffic", "car", "vehicle"
• Spatial extent: Filter on postal code
• Temporal extent: Compare parking availiblity for different events (Gentse Feesten, Winter Market, Summer Holidays)
• Filter on vehicle-type (car, bike, electric car)

8. Expected Output/Results:

• The output should be a list of datasets, sorted based on their relevancy / match with the query. It should be highlighted which aspects of the datasets match with the query. For example:
  • Matched:
    • postal code: 9000 - 9999
    • keywords: "parking", "occupancy"
    • time range: 01/01/2024-01/01/2025
  • Not Matched:
    • key words: "availibility", "traffic", "car", "vehicle"
  • time range: 01/01/2025 - now

9. Additional Information:

• /

[pietercolpaert]

Just like #2, #1 and #4 this is a case of time series data. I think time series tasks like anomaly detection and forecasting are really interesting to study towards discoverability: what other datasets would help me forecast this time series for example.

E.G., if there are traffic measurements going up on the ringroad, possibly the parking capacity will also go down here.

We’re working on an RDF vocabulary to describe time series snippets here: https://knowledgeonwebscale.github.io/rdf-timeseriessnippets/

[thcarsten]

Please find attached an example of a Linked Data Event Stream, to which I have added a SHACL description of its members using the dc:conformsTo property. The SHACL shape is not exhaustive, but it includes a few example properties.

For this example, I used the Telraam LDES, which relates to traffic data. I was unable to find an LDES related to parking data, so I used this traffic-related dataset instead. More information about Telraam can be found here: https://telraam.net/en/our-traffic-counter.

LDES telraam conformsTo.txt