Geospatial Information Technologies

Course Information
Courses and exams
Prof	Van Orshoven Jos; Lhermitte Stef; Vanden Berghe Ingrid
Courses	Lectures
Examination	Partial or continuous assessment with (final) exam during the examination period
Background
Credits	6
When?	1st semester
ECTS	Link

Evalutation:

OLA I0D98a: Geospatial databases (weight 1/6):
- Written exam, Open Book
- Previously this was oral, of which we have the following "review": Van Orshoven is really nice and calm, if you don't know the answer right away he tries to help you find the answer anyway, if you don't know it at all, he explains it to you.
OLA I0D99a: Geospatial Databases and SQL: Practical (weight 2/6)
- 2 to 4 individual or group assignments to be handed in in the course of the semester according to a prescribed timing (see Toledo)
OLA I0U99a: Geospatial data infrastructures (weight 2/6)
- Written exam, since 2025 open Book, together with I0D98a
OLA I0V00A: Geospatial data infrastructures: practical (weight 1/6)
- 2 to 3 individual or group assignments to be handed in in the course of the semester according to a prescribed timing (see Toledo)

2025

January

Written exam, both exams open book

Nobody I spoke to after the exam had a feeling they could confidently answer any questions from the SDI part apart from maybe the true/false ones so dont feel bad

Geospatial Databases

Given: a dataset with discharge measurement stations, with the station name, elevation, year of start of measurements and year of end of measurements and point coordinates, with daily measurements of the parameters discharge, sediment load, superficial flow velocity, water level, and the directors responsible for the stations in different periods. You also have a river segment dataset with average bank height, line geometry, river code, river name and Strahler order (including strahler order explanation). Furthermore, you have a dataset with the 8 largest hydropower plants. The coordinates of the hydropower turbines (point geometry), the installed power generation capacity, the reservoir polygon geometry, the reservoir capacity (km3) and dam height are all available. Furthermore, you have daily measurements of the turbine's power output.

Draw a conceptual database model (entity-relationship) that can grasp this data
What real-life questions can you answer using this db-structure?
Draw a logical database model
What types of relations can you model with UML, and which of these relations are applicable where in the DB you just drew?

Spatial Data Infrastructures

Two parts, part A counted for 80% and part B for 20% (i think?)

Part A: open questions

You have a shapefile, a geotiff and a netCDF file. (20%)
1. How can you host this data on a traditional SDI?
2. How would you do that on a cloud-based platform such as Google Earth Engine?
3. What are the advantages/disadvantages of these two approaches?
4. What are the advantages/disadvantages for producers for hosting your data on a traditional SDI or a cloud-based platform such as GEE?
5. What are the advantages/disadvantages for users for hosting the data on a traditional SDI or a cloud-based platform such as GEE?
Compare CSW and STAC in SDI context. What are the respective advantages and disadvantages in the discovery, management and access to metadata? Give example scenarios where you would prefer either one. (20%)
You want to assess the evolution of NDVI of areas in the USA hit by forest fires using Landsat imagery since 1980. How would you do this using a openEO datacube? Write out pseudocode. (20%)
You have a .zar file with climate data for a large area in n dimensions containing variables such as temperature, pressure, precipitation, etc. You want to calculate some kind of yearly weather statistic, but in a way allowing for lazy computation. Integrate this all in an interactive Leaflet map. Write this all out in pseudocode. (20%)

Part B: true/false. You only get 0.5/4 for the question if you only get the true/false right without justification, with justification this can climb to 4/4. Each question here counted for 4%

The primary goal of an SDI is to store data efficiently within an organisation.
Chunking is not relevant for local data storage as it does not impact storage efficiency.
XYZ tiling is performant and caching-friendly in applications with real-time data updates
Data regularisation is only relevant for spatial data and does not have to be applied when using temporal data
abc

2022

January

SPATIAL DATA INFRASTRUCTURES

1 open question: SWE figure given, explain all components

10 true/false questions (+ explanation):

The NGI is a geobroker outside in for the government
The main steps of data harmonisation are ... The harmonisation process can always be called standardisation
the GRB and open street map are authentic sources of data
GetFeatureInfo request of WMS given, is this correct?
Discovery, Exploration, Exploitation: availability of metadata: which one?
Geometry, topology and semantics --> CityGML, KML and ArcGis --> do they represent the first three terms
Are these measures good for measuring impedence of an SDI? (Price, Legal, transfer method, restrictions and need for preparation)
...

GEOSPATIAL DATABASES

(written because of covid)

Dataset 1:

A certain amount (I don't remember exactly) of charging poles for electrical cars or bikes with point coordinates, name, type of usage (car, bike, car&bike), type of operator (public, private), start date, date of last maintenance, type of electricity generation (solar, wind, fossil), simultanuous service capacity, ... (maybe some other things). Also for each pole the name and professional details of the operator who helped construct the pole, per sub-period were given. There are also daily statistics of the number of bikes, the amount of electricity per bike, the number of cars and the amount of electricity per car.
Certain amount of repair facilities with point coordinates, type of usage (bike, car, bike&car), opening hours, ... and the repairs done each day again with the type, time, ...
Network of roads and bike tracks with line geometry and type of road.

Questions:

Make a conceptual entity-relation diagram
Which real-life questions can be answered using this conceptual model and which GIS tools would you use to solve these questions?
Create a logical data model with help of the previously made entity-relationship model.
Is this a hybrid gDB? if not, how can you adapt it to a hybrid gDB?

2021

January

Spatial Data Infrastructures

1 open question: SWE figure given, explain all components

10 true/false questions (+ explanation):

The NGI is a geobroker outside in for the government
The main steps of data harmonisation are ... The harmonisation process can always be called standardisation
the GRB and open street map are authentic sources of data
getfeature request of WFS given. Is this correct? What is missing?
GML, CityGML and HTML are all XML-based.
SDI governance is about assessing the performance of the SDI.

Geospatial Databases

(written because of covid)

A lot of information is given about measurement stations (and measurements) and power plants on a river network

create a conceptual database model
Which questions can be answered using this conceptual model?
Create a logical data model
Which are the 4 non-spatial relationships in UML and how can you use these between the classes in your logical data model?

2019

January

Geospatial Databases

Dataset:

Example of a river network. there are different measurement stations that returns measurements like water depth, sediment load, surface speed, start of operation, end of operation... Then there are directors with a name, details... that are in charge of measurement stations, can change over time. The table of stations also contains stations that replaced older stations. Then you have data of Hydro power plants with the power output, which company manages it, a geometry of the power plant, a geometry of the basin related to it and again measurements of power output for certain time intervals. Also given is the geometry of the river with more than 20.000 segments.

Questions:

Make a logical structure (entity/relation model) of this data (expect to give some cardinalities)
What answers could you derive from this dataset? Could you measure effects of climate change? You have only 200 measurements but more than 20.000 river segments, are those useful? ("You should do some interpolation")
The table of stations also contains stations that replace older stations, how is this fitted in the database (kinship relation)
UML defines more types of relationships besides association, would some be applicable here?
How could this database be turned into a hybrid database?
Draw a logical and hybrid database based on the UML principles.
How could spatial indexes be used and implemented in this database?

2017

January

Geospatial databases

Same question as previous years, you randomly take a dataset (I think there were three different ones)

Dataset 1:

A certain amount (I don't remember exactly) of charging poles for electrical cars or bikes with point coordinates, name, type of usage (car, bike, car&bike), type of operator (public, private), start date, date of last maintenance, type of electricity generation (solar, wind, fossil), simultanuous service capacity, ... (maybe some other things). Also for each pole the name and professional details of the operator who helped construct the pole, per sub-period were given. There are also daily statistics of the number of bikes, the amount of electricity per bike, the number of cars and the amount of electricity per car.
Certain amount of repair facilities with point coordinates, type of usage (bike, car, bike&car), ... maybe something else.
Network of roads and bike tracks with line geometry and type of road.

Questions (with some answers):

spatial relationships: typical network (distance relation)
questions to answer with this dataset
make a conceptual entity-relation diagram (entity classes were: charging poles, operators, repair facilities, road-segments) specify the cardinalities
ORDB:
- the tables (logical modelling), tables for the entity classes and pivot tables for the many-to-many relationships (charging pole + operator): specify primary and foreign keys
- Lookup tables for the attributes (because numbers in the entity table are more compact than strings)
- hybrid geo-relational database:
  - this means taking the geometry out of the object-tables and storing these in a separate file. The geometry file and the object-table have a 1-1 cardinality
How to use indexing, both attribute and spatial, for the repair facilities?
- attribute index = B-tree (draw example index table and explain to him what this is + the cardinality of this table with the entity table)
- spatial index = grid, quad-tree or R-tree and then he asks to explain one or more of these

Same dataset as above: (question 1,2,3 and 4 were the same)

5.

a) What do you need to change to convert the model made in previous questions into UML.

b) Can you apply a kinship relationship to the nearest charging point?

6. apply versioning on the road network.

2016

June

Spatial Data Infrastructures

2 big questions, each on 5 points:

Explain a WMS
SWE image: explain all components (he asks this all the time)
CityGML: appearance?
GML Linestring: you have to specify what each part does. Then: you have to transform the given GML data into one that is representing a doughnut.
! Define two strings & define the outer & inner boundary. --> You should know the tags of inner & outer boundary.

True/false questions (10 questions, each on 2 points: 0.5 for the T/F, 1.5 for the explanation you provide for it).

Explain 10 (!) abbreviations & if they're correct or not
Figure: is this the right order in which a request is posed in a metadata catalog?
Is the main reason that SDI uses object-relational databases because it allows multi-user functionality? FALSE. It's one of the reasons, but other reasons = e.g. make relations between datasets.
HTTP & GML are both XML based
Table & XML translation of it: correct or not?
Are open street map & GRB authentic data?
Discovery, Exploration, Exploitation: availability of metadata: which one? (it's discovery, the statement says its exploitation).
Data transformation (of reference data) --> E.g. making it INSPIRE conform: what are the main steps? Target Schema matching, mapping & transformation?
Conflation: put two datasets together & use them.

2015

June

Spatial Data Infrastructures

Open questions: you get two of those

e-Government, explain all
the figure of SWE architecture was given. Explain all the components. Where can this technology be useful in the future? Give an example that we did not see in the lectures. Side questions: would a SWE technology be usful for the smoke detectors in this building? What would be missing?
an example of a link to open a WMS was given (from slide 20 the second one). Explain this. What is a WMS? What does 'getFeatureInfo' do? Why is no reference system given? What is BBOX? How is the image delivered?
given was the semantic UML-schema of a building in cityGML. Explain the diagram. How is the 3D nature of this model apparant in this schema? How is topology apparant in this schema?

correct/false (you get 10 of these and have to explain)

"in SDI, metadata needs to allow a user to find data, see what it's contents are and see how this data can be obtained. This type of metadata is called discovery metadata."
"The data quality elements are: completeness, positional accuracy, lineage, and usability."
"html and gml both use xml"
given: figure from cityGML which is also in the course, of a building which had rooms. "This is LoD4"
given was the xml-code of the parcels from linkvitt. But he changed it a bit. it went: <myDatabase> <Owner> <Owner> Lotte </Owner> <Parcel> Berga1:1 </Parcel> <Area>2.1</Area></Owner>. "This is a correct way of representing this table."
"An RTK service is comparable to a SOS
Given was a UML diagram. "This is a diagram that shows how metadata catalogues work."
"The properties of open data are: (I don't remember but he gave five of them)"
.

2014

January

Geospatial Databases

*It's a long question, I do not know he details

given is following data on a public transport network in Antwerp:

all (4000+) stops: coördinates, name, date of last maintenance, comfort level (shelter or open air), start of operation date, end of operation date.
342 trajectories (vector line files) with: geometry, trajectory code, trajectory name, direction (so one line can be stored twice: from A to B and from B to A, if this trajectory is indeed in both directions). It also includes:
the compagny that regulates the trajectory in a given time period (the last ten years), and all the details about this compagny regulations (price, director, ...)
yearly statistics about the number of people using the trajectory, the number of people using it without valid ticket, and the number of criminal acts on the trajectory, for the last 15 years
A road and rail map of Belgium, containing all roads, railways, tramways (over a million vectors).

now:

what spatial relations can be found in this dataset?
what questions can be answered using this dataset?
make an entity-relation diagram.
.
- transform this to an object-relational database (logical modelling)
- transform this to a geo-relational database
how can you use indexing here? illustrate using the 'Stops' entity. Do this non-spatial and spatial.

questions during oral: he always asks the cardinality, of every relation. Also, he like 'lookup-tables', ...

time is limited!

Previous years

less certainty:

given: some data

what questions can be answered with this database?
give the topological relations
make an entity-relationship diagram
convert to:
- object-relational database
- geo-relational database
do indexing for the parcels, for attributes as well as spatial.

He asked this for a dataset on farms one year. The other year, it was 121 measuring stations at flemish rivers. You get the hydrological data, and who was director of the station through time. you also get a line vector dataset with the rivers, a poly dataset with the catchments (12 polygons, 11 catchments so catchment consisted of 2 polygons...)