Topics

Competency Questions #ontology #rdf


Matteo Lissandrini (AAU)
 

Hi all,
an important initial step to align our intentions and clarify the scope of the ontology and the goals of the database is to identify a set of "Competency Questions".
Those are questions that should be addressable by the database.

Below you will find a small set taken from the literature I've found. I would ask you to think of questions that have very high importance to you.
Please accompany them with an example.
In this phase I would ask you NOT to comment on other people questions.
Yet, feel free to propose a new question and mark it as alternative to a previously proposed question.
I'm using a very naive encoding here to identify questions.


Thanks,
Matteo

(ML_Q1) Is the flow $x$ a determining product for the activity $y$ ?
          (e.g., electricity from a power plant)

(ML_Q2) Is input flow $x$ required for activity $y$?
          (e.g., coal for electricity from a power plant)

(ML_Q3) What is the amount of flow $x$ emitted as output during the time period $y$ ?
         (e.g., the emission of landfill gas)

(ML_Q4) What is the location of the agent performing the activity $y$?
         (e.g.,where is the coal power plant located)

(ML_Q5)  What other agents performing the same type of activity of agent $z$ are present in the same location $w$ ?
        (e.g, power plants in Germany)


Agneta
 

Hi Matteo

Good questions!

I would like to add:

(AG_Q1) How are the flow objects quantified/ Which units of measure are used?
                (e.g. in monetary units, weight, etc)
(AG_Q2) Does the input and output of flow objects follow mass balance?
                (e.g. iron ore and coal input used to produce steel, blast furnace gas, carbon dioxide)
(AG_Q3) Is output flow required for activity of type 'stock'?
               (e.g. output of a landfill)

comment- power plant would be an agent, electricity generation is activity, electricity is the flow object. 


Bo Weidema
 

I would like to attack this a bit more systematically:

 

Querying the core database alone (noting that the core observations/datapoints are “flows”):

 (BW_Q1) What is the direct input flow of flow-object F to activity A measured by flow-property P in location L in the time period T under macro-economic scenario S?

(example: What is the direct input of coal to steel production measured by dry mass in Germany in year 2020 under the Business-as-Usual scenario?)

(BW_Q2) What is the direct output flow of flow-object F from activity A measured by flow-property P in location L in the time period T under macro-economic scenario S?

(example: What is the direct output of steel from steel production measured by EUR2011 nominal value in Germany in year 2020 under the Business-as-Usual scenario?)

(BW_Q3) What is the determining flow of activity A in location L in the time period T under macro-economic scenario S?

(example: What is the determining flow of soybean mills in Brazil in year 2020 under the Business-as-Usual scenario?)

(BW_Q4) What is the difference between input flows and output flows of flow-property P for activity A in location L in the time period T under macro-economic scenario S?

            (example: What is the difference between input flows and output flows of dry mass for all activities globally in year 2020 under the Business-as-Usual scenario?)

 

Querying the core database together with a system model algorithm (for creating product footprints):

(BW_Q5) What is the additional input flow of flow-object F to activity A measured by flow-property P in location M in the time period T resulting from a demand of flow-object G in location M in the time period U, all under macro-economic scenario S?

(example: What is the additional input of surface water to steel production measured by wet mass in Germany in year 2020 resulting from a demand of 100 square meter of office building in Spain in year 2019 under the Business-as-Usual scenario?)

(BW_Q6) What is the additional output flow of flow-object F from activity A measured by flow-property P in location M in the time period T resulting from a demand of flow-object G in location M in the time period U, all under macro-economic scenario S?

(example: What is the additional radiative forcing from the atmospheric energy transfer measured by power per area globally in year 2020 resulting from a demand of 100 square meter of office building in Spain in year 2019 under the Business-as-Usual scenario?)

 

I will break the rule of not commenting, by pointing out that:

(ML_Q1) is an instance of (BW_Q3)
(ML_Q2) is an instance of (BW_Q5)
(ML_Q3) is an instance of (BW_Q2)
(ML_Q4) and (ML_Q5) are queries for the location of specific agents or activities, and thus possible and relevant queries, but not related to the flows that is the core observations in the database

(AG_Q1) can be obtained as an instance of either of (BW_Q1), (BW_Q2), (BW_Q5) and (BW_Q6) (AG_Q2) is similar to (BW_Q4)
(AG_Q3) is an instance of (BW_Q6)


Massimo Pizzol
 

Everything looks fine to me, I just miss some questions about impacts (and I realize that I am not sure if characterisation factors are included in our schema as properties or how …?). Examples:

 

(MP_Q1) What is the direct output flow of flow-object F from activity A measured by flow-property P in location L in the time period T under macro-economic scenario S?  

(example: What is the direct output of carbon dioxide from steel production measured in carbon dioxide equivalents in Germany in year 2020 under the Business-as-Usual scenario?) à Is this just another instance of type BW_Q1?

 

(MP_Q2) What is the additional output flow of flow-object F from activity A measured by flow-property P in location M in the time period T resulting from a demand of flow-object G in location M in the time period U, all under macro-economic scenario S?

(example: What is the additional global warming potential from steel production measured in carbon dioxide equivalents in year 2020 resulting from a demand of 100 square meter of office building in Spain in year 2019 under the Business-as-Usual scenario?) à Is this just another instance of type BW_Q6?

 

BR
Massimo

 

From: <hackathon2019@bonsai.groups.io> on behalf of "Bo Weidema via Groups.Io" <bo.weidema@...>
Reply-To: "hackathon2019@bonsai.groups.io" <hackathon2019@bonsai.groups.io>
Date: Wednesday, 13 March 2019 at 16.03
To: "hackathon2019@bonsai.groups.io" <hackathon2019@bonsai.groups.io>
Subject: Re: [hackathon2019] Competency Questions #ontology #rdf

 

I would like to attack this a bit more systematically:

 

Querying the core database alone (noting that the core observations/datapoints are “flows”):

(BW_Q1) What is the direct input flow of flow-object F to activity A measured by flow-property P in location L in the time period T under macro-economic scenario S?

(example: What is the direct input of coal to steel production measured by dry mass in Germany in year 2020 under the Business-as-Usual scenario?)

(BW_Q2) What is the direct output flow of flow-object F from activity A measured by flow-property P in location L in the time period T under macro-economic scenario S?

(example: What is the direct output of steel from steel production measured by EUR2011 nominal value in Germany in year 2020 under the Business-as-Usual scenario?)

(BW_Q3) What is the determining flow of activity A in location L in the time period T under macro-economic scenario S?

(example: What is the determining flow of soybean mills in Brazil in year 2020 under the Business-as-Usual scenario?)

(BW_Q4) What is the difference between input flows and output flows of flow-property P for activity A in location L in the time period T under macro-economic scenario S?

                (example: What is the difference between input flows and output flows of dry mass for all activities globally in year 2020 under the Business-as-Usual scenario?)

 

Querying the core database together with a system model algorithm (for creating product footprints):

(BW_Q5) What is the additional input flow of flow-object F to activity A measured by flow-property P in location M in the time period T resulting from a demand of flow-object G in location M in the time period U, all under macro-economic scenario S?

(example: What is the additional input of surface water to steel production measured by wet mass in Germany in year 2020 resulting from a demand of 100 square meter of office building in Spain in year 2019 under the Business-as-Usual scenario?)

(BW_Q6) What is the additional output flow of flow-object F from activity A measured by flow-property P in location M in the time period T resulting from a demand of flow-object G in location M in the time period U, all under macro-economic scenario S?

(example: What is the additional radiative forcing from the atmospheric energy transfer measured by power per area globally in year 2020 resulting from a demand of 100 square meter of office building in Spain in year 2019 under the Business-as-Usual scenario?)

 

I will break the rule of not commenting, by pointing out that:

(ML_Q1) is an instance of (BW_Q3)
(ML_Q2) is an instance of (BW_Q5)
(ML_Q3) is an instance of (BW_Q2)
(ML_Q4) and (ML_Q5) are queries for the location of specific agents or activities, and thus possible and relevant queries, but not related to the flows that is the core observations in the database

(AG_Q1) can be obtained as an instance of either of (BW_Q1), (BW_Q2), (BW_Q5) and (BW_Q6) (AG_Q2) is similar to (BW_Q4)
(AG_Q3) is an instance of (BW_Q6)


Bo Weidema
 

Den 2019-03-13 kl. 16.28 skrev Massimo Pizzol:

Everything looks fine to me, I just miss some questions about impacts (and I realize that I am not sure if characterisation factors are included in our schema as properties or how …?). Examples:

 

(MP_Q1) What is the direct output flow of flow-object F from activity A measured by flow-property P in location L in the time period T under macro-economic scenario S?  

(example: What is the direct output of carbon dioxide from steel production measured in carbon dioxide equivalents in Germany in year 2020 under the Business-as-Usual scenario?) à Is this just another instance of type BW_Q1?

What flows out of the steel production is carbon dioxide. 1 kg of carbon dioxide = 1 kg of carbon dioxide equivalents, so I assume what you are thinking of is rather if the output was of, say, methane, which can also be expressed in carbon dioxide equivalents under a specific macro-economic scenario, and depending on where in the impact pathway you define the equivalence. So the most consistent way of dealing with this is to follow the methane through the following ecosystem activities "atmospheric energy balance" and "temperature increase" as a result of the additional input of methane. These processes will have outputs expressed in radiative forcing and temperature change, which can of course be compared to the radiative forcing and temperature change from a kg of carbon dioxide, so that you obtain the impacts of methane in carbon dioxide equivalents at the point in the impact pathway that you are interested in.

(MP_Q2) What is the additional output flow of flow-object F from activity A measured by flow-property P in location M in the time period T resulting from a demand of flow-object G in location M in the time period U, all under macro-economic scenario S?

(example: What is the additional global warming potential from steel production measured in carbon dioxide equivalents in year 2020 resulting from a demand of 100 square meter of office building in Spain in year 2019 under the Business-as-Usual scenario?) à Is this just another instance of type BW_Q6?

The global warming potential (given a specified time horizon) is a weighting factor applied to each greenhouse gas. This is most easily included in a separate weighting step. This could be done by adding the weighting as an activity, but I would suggest to use the E matrix suggested in http://lca-net.com/p/2865 for this purpose.

Best regards

Bo



Massimo Pizzol
 

Thanks.

 

Am I right that your comment is similar what you have replied already here on GDP? My impression is that like in the GDP discussion you are proposing a breakdown of each modelling step by creating an activity object for each model variable (e.g. along an impact chain). Since, however, there might be a databases/ file somewhere on the web where a list of characterisation factors (like there are GDP data per country online), this dataset on the web could be linked to our flows directly and used for e.g. for validation the results obtained from the query mentioned above, or even for calculations without using the model breakdown.

 

However, I don’t understand why in the first example you want to “follow the methane through some activities” and in the second one you want to use directly “weighting factors”. Seems inconsistent to me.

 

BR

Massimo

 

 

 

 

 

From: <hackathon2019@bonsai.groups.io> on behalf of "Bo Weidema via Groups.Io" <bo.weidema@...>
Reply-To: "hackathon2019@bonsai.groups.io" <hackathon2019@bonsai.groups.io>
Date: Wednesday, 13 March 2019 at 16.51
To: "hackathon2019@bonsai.groups.io" <hackathon2019@bonsai.groups.io>
Subject: Re: [hackathon2019] Competency Questions #ontology #rdf

 

Den 2019-03-13 kl. 16.28 skrev Massimo Pizzol:

Everything looks fine to me, I just miss some questions about impacts (and I realize that I am not sure if characterisation factors are included in our schema as properties or how …?). Examples:

 

(MP_Q1) What is the direct output flow of flow-object F from activity A measured by flow-property P in location L in the time period T under macro-economic scenario S?  

(example: What is the direct output of carbon dioxide from steel production measured in carbon dioxide equivalents in Germany in year 2020 under the Business-as-Usual scenario?) à Is this just another instance of type BW_Q1?

What flows out of the steel production is carbon dioxide. 1 kg of carbon dioxide = 1 kg of carbon dioxide equivalents, so I assume what you are thinking of is rather if the output was of, say, methane, which can also be expressed in carbon dioxide equivalents under a specific macro-economic scenario, and depending on where in the impact pathway you define the equivalence. So the most consistent way of dealing with this is to follow the methane through the following ecosystem activities "atmospheric energy balance" and "temperature increase" as a result of the additional input of methane. These processes will have outputs expressed in radiative forcing and temperature change, which can of course be compared to the radiative forcing and temperature change from a kg of carbon dioxide, so that you obtain the impacts of methane in carbon dioxide equivalents at the point in the impact pathway that you are interested in.

(MP_Q2) What is the additional output flow of flow-object F from activity A measured by flow-property P in location M in the time period T resulting from a demand of flow-object G in location M in the time period U, all under macro-economic scenario S?

(example: What is the additional global warming potential from steel production measured in carbon dioxide equivalents in year 2020 resulting from a demand of 100 square meter of office building in Spain in year 2019 under the Business-as-Usual scenario?) à Is this just another instance of type BW_Q6?

The global warming potential (given a specified time horizon) is a weighting factor applied to each greenhouse gas. This is most easily included in a separate weighting step. This could be done by adding the weighting as an activity, but I would suggest to use the E matrix suggested in http://lca-net.com/p/2865 for this purpose.

Best regards

Bo

 


Bo Weidema
 

Den 2019-03-13 kl. 17.07 skrev Massimo Pizzol:

Thanks.

 

Am I right that your comment is similar what you have replied already here on GDP?
My impression is that like in the GDP discussion you are proposing a breakdown of each modelling step by creating an activity object for each model variable (e.g. along an impact chain). Since, however, there might be a databases/ file somewhere on the web where a list of characterisation factors (like there are GDP data per country online), this dataset on the web could be linked to our flows directly and used for e.g. for validation the results obtained from the query mentioned above, or even for calculations without using the model breakdown.

Well, you can aggregate any number of activities (and thus also steps in an impact pathway) and thus obtain the desired characterisation factors (for non-LCA people: A characterisation factor is an output flow from a characterisation activity given in relation to the input of the determining flow to be characterised). And as such you can also store such aggregated activity data obtained as "raw" data, e.g. for comparison to the aggregated result of a specific more detailed modelling.

However, I don’t understand why in the first example you want to “follow the methane through some activities” and in the second one you want to use directly “weighting factors”. Seems inconsistent to me.

It stems from the fact that you asked about two different things: 1) carbon dioxide equivalents, which is a generic concept, 2) Global Warming Potentials, which are specific weighting factors for GreenHouseGases under a specific normative paradigm.

Best regards

Bo   

From: <hackathon2019@bonsai.groups.io> on behalf of "Bo Weidema via Groups.Io" <bo.weidema@...>
Reply-To: "hackathon2019@bonsai.groups.io" <hackathon2019@bonsai.groups.io>
Date: Wednesday, 13 March 2019 at 16.51
To: "hackathon2019@bonsai.groups.io" <hackathon2019@bonsai.groups.io>
Subject: Re: [hackathon2019] Competency Questions #ontology #rdf

 

Den 2019-03-13 kl. 16.28 skrev Massimo Pizzol:

Everything looks fine to me, I just miss some questions about impacts (and I realize that I am not sure if characterisation factors are included in our schema as properties or how …?). Examples:

 

(MP_Q1) What is the direct output flow of flow-object F from activity A measured by flow-property P in location L in the time period T under macro-economic scenario S?  

(example: What is the direct output of carbon dioxide from steel production measured in carbon dioxide equivalents in Germany in year 2020 under the Business-as-Usual scenario?) à Is this just another instance of type BW_Q1?

What flows out of the steel production is carbon dioxide. 1 kg of carbon dioxide = 1 kg of carbon dioxide equivalents, so I assume what you are thinking of is rather if the output was of, say, methane, which can also be expressed in carbon dioxide equivalents under a specific macro-economic scenario, and depending on where in the impact pathway you define the equivalence. So the most consistent way of dealing with this is to follow the methane through the following ecosystem activities "atmospheric energy balance" and "temperature increase" as a result of the additional input of methane. These processes will have outputs expressed in radiative forcing and temperature change, which can of course be compared to the radiative forcing and temperature change from a kg of carbon dioxide, so that you obtain the impacts of methane in carbon dioxide equivalents at the point in the impact pathway that you are interested in.

(MP_Q2) What is the additional output flow of flow-object F from activity A measured by flow-property P in location M in the time period T resulting from a demand of flow-object G in location M in the time period U, all under macro-economic scenario S?

(example: What is the additional global warming potential from steel production measured in carbon dioxide equivalents in year 2020 resulting from a demand of 100 square meter of office building in Spain in year 2019 under the Business-as-Usual scenario?) à Is this just another instance of type BW_Q6?

The global warming potential (given a specified time horizon) is a weighting factor applied to each greenhouse gas. This is most easily included in a separate weighting step. This could be done by adding the weighting as an activity, but I would suggest to use the E matrix suggested in http://lca-net.com/p/2865 for this purpose.

Best regards

Bo

 

--


Massimo Pizzol
 

>>> you asked about two different things: 1) carbon dioxide equivalents, which is a generic concept, 2) Global Warming Potentials, which are specific weighting factors for GreenHouseGases under a specific normative paradigm.

Thanks again. All clear.

(and now returning back to the original thread…) What I was asking for the is the same type of information: a value measured in CO2-equivalents. But in the first case it was this was “just” the value associated with a a specific flow (value of the carbon dioxide (or methane) emission expresses in the CO2-equivalents unit.), whereas in the second case was the value obtained from the calculation over the product system (sum of all the GHG emissions of the product system converted into the CO2-equivalents unit). I didn’t know how to formulate it correctly according to the format Matteo asked. But these would be typical LCA queries IMO.

 

Massimo

,_._,_


Groups.io Links:

You receive all messages sent to this group.

View/Reply Online (#125) | Reply To Group | Reply To Sender | Mute This Topic | New Topic
Mute #ontology | Mute #rdf
Your Subscription | Contact Group Owner | Unsubscribe [massimo@...]

_._,_._,_


mmremolona@...
 

Hi all,

Sorry for not participating as much the past few weeks. I'm trying to catch up with what everyone has said so far.

In terms of these competency questions. I guess the question that Massimo is asking is with respect to time scales and time windows. I'm not entirely familiar with the dataset that is available in the domain, but these time scales for measurements can cause some incongruity in the representation that is finally done in the ontologies. I'm not sure if the questions I ask are of the type to be included in these competency questions but my opinions are as follows:

(MR_Q1) What is the time granularity of the data that we acquire? This includes flow rates and production statistics. I also assume this varies with the different sources of data. Some data may already be averaged (Do we handle these differently?).
(MR_Q2) Are we going to aggregate data as part of the ontology specification or is this left for other parts of the pipeline? And if we are to aggregate data, to what degree and time scales? (per hour, per day, per week - I think this depends on how often we aggregate data and what data is available, I don't think a per minute data is significant enough in the overall scheme of LCA but I might be wrong) 

As of now, these are the questions that came to my head as I'm reading along the threads in this group. I'll post more ideas as I come across them.

Best,

Miguel Remolona


Bo Weidema
 

Dear Miguel,

These are relevant issues. However, for the time being, we have restricted ourselves to data that relect averages with a duration of minimum 1 year. This is because these are the typical data used, and a larger granularity would risk an overcomplication relative to the typical data in use in the domain. Nevertheless, I believe that in the future, it will be relevant to allow more flexibility here, and I think that will also be possible without actually changing the ontology. The current restriction is not ontological, just practical.

Best regards

Bo

Den 2019-03-18 kl. 05.24 skrev mmremolona via Groups.Io:

Hi all,

Sorry for not participating as much the past few weeks. I'm trying to catch up with what everyone has said so far.

In terms of these competency questions. I guess the question that Massimo is asking is with respect to time scales and time windows. I'm not entirely familiar with the dataset that is available in the domain, but these time scales for measurements can cause some incongruity in the representation that is finally done in the ontologies. I'm not sure if the questions I ask are of the type to be included in these competency questions but my opinions are as follows:

(MR_Q1) What is the time granularity of the data that we acquire? This includes flow rates and production statistics. I also assume this varies with the different sources of data. Some data may already be averaged (Do we handle these differently?).
(MR_Q2) Are we going to aggregate data as part of the ontology specification or is this left for other parts of the pipeline? And if we are to aggregate data, to what degree and time scales? (per hour, per day, per week - I think this depends on how often we aggregate data and what data is available, I don't think a per minute data is significant enough in the overall scheme of LCA but I might be wrong) 

As of now, these are the questions that came to my head as I'm reading along the threads in this group. I'll post more ideas as I come across them.

Best,

Miguel Remolona
--


Matteo Lissandrini (AAU)
 

Dear all,

I've collected the discussion and something more re: competency questions in the wiki for the RDF framework repository[1],
this will have to be restructured.
Please feel free to fix any typo or other issue you'll see.
Also, let me know if you add any new competency question.


There are a number of things still open on this. For instance, in the questions come up the concept of macroeconomic scenario which is not present in the data I've seen.
To this, probably (or maybe not, please let me know) connects the issue on Input/Output.

In our last conversation it appeared this very important detail that was missing to me (and so to the modeling):
The EXIOBASE data is a specific static snaptshot, in this data the center is actually the activities and each flow is either input to 1 or output to 1 activity only.
E.g. there is this steel production SP1 that consume some  coal C1 and outputs some steel S1.
Then there is this other steel production SP2 that consume some different coal C2 and outputs some other steel S2.
Then there is this coal mine CM1 that outputs some coal C3.
Then there is this other coal mine CM2 that outputs some coal C4.

When we reason about product footprint, then intervenes some other data/analysis process that links in some way flows, so only at this point we can record:
That the coal C4 from CM2 is actually used as the input coal C1 for SP1.
The coal C3 from CM1 is actually the input coal for SP2.

So C4=C1 -> at the same time output (of CM2) and input (for SP1), is this the issue ?

Yet, there may be cases like the following:
CM1 which outputs C3 actually split 70% in C1 for  SP1 and 30% in C2 for SP2.

Please, let me know if I'm understanding this correctly.



Thanks a lot,
Matteo















[1]https://github.com/BONSAMURAIS/BONSAI-ontology-RDF-framework/wiki/The-BONSAI-Ontology-and-RDF-Framework


 

This is a very important question! The following is my opinion, and other might have a different perspective.

Questions over what can really substitute for what (e.g. for coal this is sulfur content, energy density, but also in general lignite takes totally different handling than bituminous) are long known to be difficult questions; similarly, the correct way of modelling markets with multiple providers, trade, and re-export is also tricky. They are tricky because in most cases we have to make value judgments in what we think is the best model, without really being able to get the "right" answer.

As such, these decisions should be done by the system modelling software (see my recent blog post), and should not be addressed in the data format. We want to be able to try multiple approaches, and be able to quantify the effects of different choices. Instead, the data format should be able to represent different kinds of coal, their origin locations, trade patterns (trade is an activity, the same as other activities), and the properties of these coals. The data format can also give the volume of specific kinds of coals consumed by various activities in a region. The system model is responsible for taking this large set of data points, and creating a balanced view of a possible world.


Bo Weidema
 

Den 2019-03-18 kl. 23.35 skrev Matteo Lissandrini (AAU):

the concept of macroeconomic scenario which is not present in the data I've seen.
Please see the BONSAI glossary: https://github.com/BONSAMURAIS/bonsai/wiki/Glossary

In our last conversation it appeared this very important detail that was missing to me (and so to the modeling):
The EXIOBASE data is a specific static snaptshot, in this data the center is actually the activities and each flow is either input to 1 or output to 1 activity only.
When we reason about product footprint, then intervenes some other data/analysis process that links in some way flows, so only at this point we can record:
That the coal C4 from CM2 is actually used as the input coal C1 for SP1.
Yes, there are in fact (at least) two different instances of the database:

- one before linking, in which flows are only recorded as being either inputs to or outputs from an activity

- one after linking (which implies the application of the algorithms of a specific system model, as described by Chris), where each flow is recorded as a flow between two specific activities

The ontology can (or should be able to) handle both these instances.

Bo


 

On Tue, 19 Mar 2019 at 09:39, Bo Weidema <bo.weidema@...> wrote:

Den 2019-03-18 kl. 23.35 skrev Matteo Lissandrini (AAU):

the concept of macroeconomic scenario which is not present in the data
I've seen.
Please see the BONSAI glossary:
https://github.com/BONSAMURAIS/bonsai/wiki/Glossary

In our last conversation it appeared this very important detail that
was missing to me (and so to the modeling):
The EXIOBASE data is a specific static snaptshot, in this data the
center is actually the activities and each flow is either input to 1
or output to 1 activity only.
When we reason about product footprint, then intervenes some other
data/analysis process that links in some way flows, so only at this
point we can record:
That the coal C4 from CM2 is actually used as the input coal C1 for SP1.
Yes, there are in fact (at least) two different instances of the database:

- one before linking, in which flows are only recorded as being either
inputs to or outputs from an activity

- one after linking (which implies the application of the algorithms of
a specific system model, as described by Chris), where each flow is
recorded as a flow between two specific activities

The ontology can (or should be able to) handle both these instances.
This is, of course, correct, though it does help clarify things for me.

However, perhaps this needs a different conceptual approach? Perhaps
something like a resolved exchange, which has isInputOf *and*
isOutputOf? Probably the language would have to be adjusted. But I
guess we need this for trading activities in any case.

Bo





--
############################
Chris Mutel
Technology Assessment Group, LEA
Paul Scherrer Institut
OHSA D22
5232 Villigen PSI
Switzerland
http://chris.mutel.org
Telefon: +41 56 310 5787
############################


Matteo Lissandrini (AAU)
 

Thanks Bo for the clarification

the concept of macroeconomic scenario which is not present in the data
I've seen.
Please see the BONSAI glossary:
https://github.com/BONSAMURAIS/bonsai/wiki/Glossary

I've seen this, my concern is that the EXIOBASE data does not contain any explicit information about this, right?
So this information should come from somewhere else(?) and how this is represented in the database should be defined.
At the moment my best guess is to have a "named graph" with associated metadata (I can explain better in the call).


That the coal C4 from CM2 is actually used as the input coal C1 for SP1.
Yes, there are in fact (at least) two different instances of the database:

- one before linking, in which flows are only recorded as being either
inputs to or outputs from an activity

- one after linking (which implies the application of the algorithms of
a specific system model, as described by Chris), where each flow is
recorded as a flow between two specific activities

The ontology can (or should be able to) handle both these instances.

I think I understand this. Again this information may come from different sources (e.g., the algorithms).
We need to extend the model so this is represented correctly.

Also, I had a quick chat with Massimo so that I could better understand the technicality of input & output and how they are read.
I think this is in line to this two versions of the data as referred by Bo.

In general, I agree that the two classes (input/output) are redundant in a sense.
We could remove the classes and keep only the predicates and we will not lose information.
Note that the two classes are to distinguish the Flow, not the Flow-object (or the object that flows).
So the object that flows can flow out from an activity and into another one (here are the two flows).

The reason of the two classes is to help formulating the right query: if you ask if a flow is an input or an output, having the two classes you can ask exactly that, without the two classes you need to ask : is there an activity this flow is an input of?
You still get the same answer, but you need a different question.





Best,
Matteo



---
Matteo Lissandrini

Department of Computer Science
Aalborg University

http://people.cs.aau.dk/~matteo