Path
Data Supply
Data Supply Line
K—
Data Supply Line
Data supply defines f data arives
from a certain sensor ata certain
time with a certain reference (e.0.
tag or file name o file path), then
where does it come from, what is the
clientand whatis the project.
Data Transformation
[+———————specialise————]
Movement
Data Ingestion
Data Order
Data Ingestion Line
Data Order Line
Data Ingestion Line
Data Order Line
Data Ingestion (Lines) rack what type of
data is ingested from which sensor when
from where and in which project.
Data Order (Lines) define which
evaluations should be produced in which
project
T
causality
specialise:
Data Analysis
Data Transformation Line
specialise-
Data Analysis Line
Data Transformation Line
Data Transformation defines what Data Analysis Line
type of data and which operation is
needed 1o evaluale a certain type of
data.
Data Analysis (Lines) track which data of
which type from which source from which
sensor s analysed with which operation
using which configuration in which project
Jaestnation
Node
Organisation
Organisation represents either a
juridical organisation (a client
company, a company owning
data), ora physical site, (2
building or measrement site) or
both,
section project)
forwhich
clent.
agaregate]
Item
Resource
Data Product
Data Stream
Data Array
Sensor
A Data Stream behaves sinilar to
a big file. When a sensor ingests
raw data into Wendin it can be.
appended to a data stream
Person
Data Bucket Stream
Person represents a natural
person (user, developer, clien,
ouner of data, engineer,..)
Project
Project represents a typical project
in a service company.
Data Bucket Stream is an
alternative to Data Stream which
bridges Wendelin to a fle system
world. Each ingested block of data
is ingested as a seperate bucket
Data Bucket Stream supports
iteration by key as well as iteration
by order of ingestion. Data Bucket
stream can represent“a bunch of
files" or a complete directory.
structure.
Data Aray is the main big data
store of wendelin. Dala is stored
in an n-dimensional zbigarray
which gives a numpy compatible
Out-of-core view of array data.
‘This means that only the actually
processed data is loaded into
ram, while the zbigarray can be
bigger than ram. Zbigarray allows
to use the complete PyData
software stack ransparently with
wendelin, such as scipy, pandas
or scikitieam as well s ofher
machine leaming software such
astensrflow,
Represents a single sensor
ingesting Data into Wendelin
Data Acquisition Unit
A data product represents a certain
type of data, such as the raw
background dala of a specific type
of sensor, or the evaluated data for
a specific norm
Represents a unitwhich ingests
data into Wendelin. It can have ane
or multiple Sensors, for example a
measurement system.
Device Configuration
Device configuration represents the
configuration of a Sensor or Data
Acquisition Unit. Configuration can
Data Descriptor
be defined in PopertySheets or
JSON Schema
Progress Indicator
Data Descriptor stores
application- or project specific
metadaa felated 1o a Data
Array.
Data Operation
A data operation represents a
certain type of operation on daa
©.g.converting raw data flom a
data stream 10 a data array,
resampling data or evaluating
data. It also defines which scriptis
used to perform the actual
operation.
Progress Indicator represents the
processing state of a Data Analys:
Ittells how much of the input data
has been already processed (e.g
converted or evaluated) into
outputdata
Data Configuration
Data Configuration represents
the configuration used for a data
analysis. Configuration can be
defined in PopertySheets or
JSON Schema