Sabrina: Modeling and Visualization of Economy Data with Incremental Domain Knowledge

Alessio Arleo; Chao Jia; Christos Tsigkanos; Ilir Murturi; Johannes Sorger; Manfred Klaffenboeck; Michael Wimmer; Roger A. Leite; Schahram Dustdar; Silvia Miksch

REVIEW 3 major objections 6 minor 26 references

Sabrina: Modeling and Visualization of Economy Data with Incremental Domain Knowledge

T0 review · 3 major / 6 minor · reviewed 2026-08-14 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read Sabrina infers firm-to-firm financial transaction networks from public micro data, macro sector flows, and incremental expert knowledge, and visualizes them for economic analysis.

desk verdict A genuine design study with an honest limitation statement, but the usefulness claim outruns the evidence because the user study never checks model credibility. read the letter →

arxiv 1908.07479 v2 pith:KQCNEERW submitted 2019-08-05 q-fin.GN cs.HC

Alessio Arleo , Christos Tsigkanos , Chao Jia , Roger A. Leite , Ilir Murturi , Manfred Klaffenboeck , Schahram Dustdar , Michael Wimmer

show 2 more authors

Silvia Miksch Johannes Sorger

This is my paper · ORCID

classification q-fin.GNcs.HC

keywords visualanalyticsfinancialtransactionnetworksconstraintsatisfactionsatisfiabilitymodulotheoriesdomainknowledgeinfusioninput-outputtablesgeo-spatialvisualizationfirm-to-firmflows

verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

The paper tries to solve a data gap: monetary flows between individual firms are generally not public, even though information about the firms themselves and about sector-level flows is available. It claims that these missing firm-to-firm transactions can be recovered as a constraint-satisfaction problem, in which public micro data, high-level input-output tables, and incremental expert insights are encoded as constraints, and any satisfying assignment becomes a weighted directed transaction graph. Sabrina is the visual-analytics system built around that pipeline, letting analysts view firms on hexagonal region grids, follow transaction arcs between regions, and switch between co-existing models over time. If correct, this gives financial analysts a way to explore inter-firm trade relations, monitor regional and sectoral change, and test what-if policy scenarios without proprietary transaction data.

What carries the argument

The load-bearing mechanism is the SMT (satisfiability modulo theories) constraint model of transactions. Domain bounds, high-level macro flows, and expert knowledge are written as first-order constraints with quantifiers over finite firm sets and integer linear arithmetic; an SMT solver computes a satisfying assignment, and that assignment is decoded into the edge weights of a weighted directed firm-to-firm graph. The visual interface carries the analysis onward: dynamic hexagonal binning of firm locations addresses overplotting, while 2.5D flow-map arcs encode transaction direction and amount and reduce occlusion, with a time slider and model-switching menu supporting temporal and comparative tasks.

What would settle it

Take a region where true firm-to-firm transaction data are known from a proprietary source, such as bank transfer records or tax filings, run the Sabrina pipeline using only the public micro and macro constraints, and compare the inferred graph to the ground truth. If the inferred edges and weights show little or no overlap with actual transactions even when substantial domain knowledge is added, the claim that satisfying assignments approximate reality is refuted; if they track the true flows, the approach is supported.

Watch

Extended reading notes

Core claim

On the paper's own terms, the central discovery is a pipeline that turns heterogeneous financial data into an explorable firm-level transaction model. Ground truth about individual firms (location, sector, cash-flow figures) and publicly known macroscopic flows, such as sector-to-sector monetary outflows from input-output tables, are combined with domain knowledge supplied by the analyst. The combination is encoded as first-order logical formulas over finite sets with integer linear arithmetic, and any valuation satisfying the formula is declared a valid model; the more constraints are added, the closer the model is claimed to come to reality. The resulting model is a weighted directed graph whose edges carry estimated transaction amounts, and the Sabrina interface renders these graphs on a map with hexagonal binning and flow-map arcs, supporting five analyst tasks from status overview to model comparison. A qualitative user study with three domain experts is offered as evidence that the tasks are fulfilled and that the analysis process is eased.

Load-bearing premise

The load-bearing premise is that any assignment of firm-to-firm flows that satisfies the stated constraints is a meaningful approximation of real transactions; because many graphs can satisfy the same constraints, a model built from only a few constraints may visualize edges that are arbitrary artifacts rather than economic facts.

Editorial extensions

If this is right

Analysts can inspect inferred inter-firm monetary flows at several aggregation levels, from individual firm to region and sector, without owning proprietary transaction data.
Domain knowledge can be added incrementally, so the analyst can refine a model or run what-if scenarios by tightening constraints and regenerating the graph.
Multiple transaction models can coexist and be compared, supporting task T4 and enabling hypothesis testing about which constraint sets yield plausible regional structures.
The time dimension lets users monitor how firm density and aggregate cash flow change year to year, which can be used to ask whether policy changes are reflected in firm-level outcomes.
Because the constraint encoding is independent of any particular financial modeling technique, the same system can integrate outputs from other model-building approaches as expert knowledge.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

My inference: if the inference approach proves reliable, the same constraint-satisfaction pipeline could reconstruct other missing pairwise networks, such as inter-bank lending, supply-chain links, or ownership networks, where only aggregate flows and node attributes are observed.
My inference: the underdetermination noted in Section 3.2 means that a visualized edge may be an artifact of the constraint set; a natural testable extension is to validate inferred edges against a small proprietary dataset of true transactions and measure edge overlap or weight correlation.
My inference: the design choices for geo-spatial network visualization, hexagonal binning plus 2.5D flow-map arcs, transfer to other domains with node-link data on maps, independent of the financial context.

Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Desk Editor's Note

read the letter

Short version: this is a legit design study with a clear visual system, but the evidence for the usefulness claim is thinner than the abstract implies, because the user study never checks whether the inferred transaction networks resemble reality. I'd send it to a VA venue with a request to tighten claims and add a model-credibility check.

What's new: the paper's contribution is integrating an SMT-based transaction inference pipeline (developed in the companion paper [25]) into a geospatial visual analytics interface for macro/micro economic data. That integration, the task set T1–T5, and the three-expert evaluation are genuinely new relative to the cited work. The visualization design is thoughtful: hexagonal binning for firm density, 2.5D flow maps for transactions, time navigation, and model comparison. The design rationale is honest, and the authors explicitly state the underdetermination limitation in Section 3.2. That is good practice.

Where it's soft: the evaluation is a scripted demo with three self-selected experts from one organization, no baseline, no quantitative metrics. That is common for short papers, but the abstract claims Sabrina 'eases the analysis process' and the user study does not really test that. Phase 3 asks about features and impressions, not about whether the inferred edges are plausible. The stress-test note is right that the underdetermination is not bounded. The paper does not say how many constraints are needed for the approximation to be meaningful, nor does it enumerate the constraints used in the Austrian case study. Since the transaction data are unavailable, the experts are approving the interface, not the model. This matters most for tasks T3–T5. A reader can't tell if the visualized edges are economic signals or arbitrary satisfying assignments.

No code or data are shipped, and validation of the inference itself is delegated to [25]. That is acceptable for a prototype paper, but it means the model's credibility is not independently checkable here. The runtime of 157 minutes for 500 firms is reported, but there is no scaling analysis.

The math is not circular: constraints come from external macro data, and SMT finds a satisfying assignment. The citation pattern seems fair, with [25] carrying the modeling load. This paper is a reasonable design study worth engaging, and I would not cite it for the inference method — I'd cite it for the visualization design if at all.

My recommendation: send it to peer review at a visualization venue. It needs a modest revision: moderate the abstract's usefulness claim, add a paragraph on the constraint-to-plausibility gap, and ideally include a small sanity check of inferred edges against known facts. But it is a real paper with a real prototype and honest limitations.

Referee Report

3 major / 6 minor

Summary. The paper presents Sabrina, a visual analytics prototype for exploring Austrian firm-level economic data. The system combines micro data (firm locations, sectors, financial figures) with macro data (sector-to-sector transaction flows) and user-supplied domain knowledge, encodes the available information as first-order constraints within satisfiability modulo theories (SMT), and uses the CVC4 solver to produce a weighted directed firm-to-firm transaction network. The visualization provides a hexagonal aggregation view, a flow-map view of inferred transactions, time navigation, filtering, and switching among multiple transaction models. A qualitative user study with three experts from the Austrian Chamber of Commerce reports positive feedback on the overview, aggregation, time-based filtering, and interest in the inferred transaction relations. The authors frame the work as a design study and an exploratory validation of the Sabrina approach.

Significance. If the model-inference step is accepted, Sabrina offers a plausible way to turn heterogeneous public data into a visually explorable transaction network, and the visualization design responds to real problems of overplotting and edge-node occlusion. The SMT-based formulation is clean and does not depend on fitted parameters; the authors also explicitly acknowledge that the constraint-satisfaction approach is underdetermined. However, the paper's central usefulness claim rests on a bridge between an arbitrary satisfying assignment and economic reality, and the evaluation never validates that bridge. The significance is therefore conditional: the paper demonstrates a potentially useful visual interface, but the evidence for the economic meaningfulness of the inferred transaction network, and thus for tasks T3-T5, is currently missing.

major comments (3)

[Section 3.2] The underdetermination of the inferred transaction network is acknowledged but never bounded, and the evaluation in Section 4 does not test whether the visualized edges carry economic meaning. The statement "there can be many different graph variants satisfying the same constraints; the more constraints are specified, the better the model approximates reality" is load-bearing for the usefulness claim: if the satisfying assignment is arbitrary, then the network views in Fig. 1b are not evidence about the economy. The paper provides no constraint sets from the Austrian case study, no sensitivity analysis across different satisfying assignments, and no comparison of inferred edge statistics to known regional or sector-level flows beyond the macro constraints already encoded. Since tasks T3 and T4 depend directly on transaction edges, the experts' positive reaction to "potential business relations" cannot be separated from approval of the interface. The manuscript should either demonstrate that typical constraint sets produce stable edge-level statistics, or explicitly limit the usability claim to the interface and treat the network content as a hypothesis-generation device rather than a validated model.
[Section 3.1 and Section 4] Task T5, "Infusing domain knowledge," is listed as a core task and emphasized in the abstract and introduction, but the prototype does not actually support it. Section 4 reports that "the model specifications are currently scripted outside of the visual interface" and that the experts "expressed interest in a way to visually modify the constraints... a feature that is not yet supported by our prototype." Thus the system as presented cannot perform T5, and the evaluation only captures an expressed desire, not evidence of use. The authors should either remove T5 from the supported task list, describe a concrete interface mechanism for constraint editing, or clearly label T5 as future work rather than as a task fulfilled by Sabrina.
[Section 4] The user study is a two-hour qualitative session with three self-selected experts, a scripted demo, a free-exploration period, and verbal feedback. There is no baseline condition, no task-completion measure, no interaction log analysis, and no explicit question about whether the experts believe the inferred transaction edges are plausible or whether the network view changed any analytical conclusion. Given the central claim that Sabrina "eases their analysis process," this evidence is weak. It is acceptable as a preliminary, exploratory design validation only if the claims are softened accordingly and the underdetermination caveat from Section 3.2 is carried into the interpretation of the results. At minimum, the evaluation should report the protocol questions verbatim and distinguish between feedback on visual features and feedback on the trustworthiness of the inferred model.

minor comments (6)

[References] References [23] and [24] are the same paper and should be merged; the first occurrence lacks a DOI.
[Section 3.2] The phrase "monetary in&out flows" should read "monetary in/out flows" for clarity.
[Figure 1] The figure caption lists items (a) through (g), but the text in Section 3.3 refers to "details panel" and "configuration pane" without consistently matching the labeled panels (e) and (f); please align the descriptions with the figure.
[Section 3.2] The system name "Sabrina" and the data source "Sabina" are visually and phonetically similar and may confuse readers; consider adding an explicit note distinguishing the system from the database.
[Section 4] The recruitment of experts "because of their interest in Sabrina and their compatibility with the scope of the project" should be described as a convenience sample, and the implications for generalizability should be acknowledged.
[Section 3.2] The reported runtime of 157 minutes for 500 firms should be contextualized: it is unclear whether this is for a single model, how the runtime scales with the number of constraints, and whether this is acceptable for the intended interactive workflow.

Circularity Check

0 steps flagged · score 0.0 of 10

No circular derivation: transaction graphs are explicitly defined as constraint-satisfying assignments, and the usefulness claim rests on a user study rather than on the self-cited companion paper.

full rationale

The pipeline's only derivation-like step is the SMT encoding of macro data, domain bounds, and expert knowledge into first-order formulas; any satisfying valuation is output as the transaction graph. This is definitional constraint satisfaction, not a prediction fitted to a target, so no fitted-input-called-prediction cycle exists. The paper explicitly acknowledges the model's underdetermination: 'there can be many different graph variants satisfying the same constraints; the more constraints are specified, the better the model approximates reality. This is an inherent limitation of the approach.' That is an evidentiary/validity limitation, not a circular reduction: the output is not defined in terms of the claim it is used to support. Reference [25] is a self-citation by overlapping authors, but it is used only as a pointer for a more extensive analysis of the modeling process, not as the source of the constraints or equations in this paper, and it is not invoked as a uniqueness theorem; the usefulness claim is supported by the three-expert user study. The user study asks experts about interface features and task support rather than verifying inferred edges, which is a weakness in external validation, not circularity. There are no equations or definitions that make a predicted quantity equal to an input by construction.

Assumptions & free parameters 0 free parameters · 3 assumptions · 0 invented entities

The central claim rests on assumptions that macro tables and expert statements can constrain firm-level transactions, and that a satisfying SMT assignment is meaningful. No free parameters are fitted and no invented entities are introduced.

assumptions (3)

domain assumption Macroeconomic sector-to-sector flow tables (IO tables) are reliable ground truth for aggregate firm behavior.
The pipeline uses these tables as constraints to infer firm transactions, described in Section 3.2 under High-Level Behavior. If the IO tables are inaccurate, the inferred network is invalid.
domain assumption A satisfying assignment of the SMT formula approximates real firm-to-firm transactions.
The paper acknowledges in Section 3.2 that many graphs satisfy the same constraints and that more constraints improve approximation. This is an unverified assumption on which the usefulness of the system rests.
domain assumption Experts' stated domain knowledge is accurate and correctly encoded as constraints.
Section 3.2 Expert Knowledge allows users to introduce constraints such as trading limits, but there is no mechanism in the paper to validate those statements against reality.

how reviews work

0 comments

Cite this review

Pith. "Pith review of Sabrina: Modeling and Visualization of Economy Data with Incremental Domain Knowledge." pith.science (2026). https://pith.science/paper/KQCNEERW

@misc{pith2026190807479,
  author       = {Pith},
  title        = {Pith review of: Sabrina: Modeling and Visualization of Economy Data with Incremental Domain Knowledge},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/KQCNEERW}},
  note         = {Machine review of arXiv:1908.07479}
}

read the original abstract

Investment planning requires knowledge of the financial landscape on a large scale, both in terms of geo-spatial and industry sector distribution. There is plenty of data available, but it is scattered across heterogeneous sources (newspapers, open data, etc.), which makes it difficult for financial analysts to understand the big picture. In this paper, we present Sabrina, a financial data analysis and visualization approach that incorporates a pipeline for the generation of firm-to-firm financial transaction networks. The pipeline is capable of fusing the ground truth on individual firms in a region with (incremental) domain knowledge on general macroscopic aspects of the economy. Sabrina unites these heterogeneous data sources within a uniform visual interface that enables the visual analysis process. In a user study with three domain experts, we illustrate the usefulness of Sabrina, which eases their analysis process.

Figures

Figures reproduced from arXiv: 1908.07479 by the authors.

**Figure 2.** Sabrina data pipeline. The ground truth consisting of micro [PITH_FULL_IMAGE:figures/full_fig_p002_2.png] view at source ↗

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

26 extracted references · 23 canonical work pages

[25]

Tsigkanos, A

C. Tsigkanos, A. Arleo, J. Sorger, and S. Dustdar. How do ﬁrms trans- act? Guesstimation and validation of ﬁnancial transaction networks with satisﬁability. In IEEE 20th International Conference on Informa- tion Reuse and Integration, Los Angeles, California, USA, July 30 - August 1 2019, (to appear)

work page 2019
[1]

https://github.com/keplergl/kepler.gl

kepler.gl. https://github.com/keplergl/kepler.gl. Accessed: 2019-06-12

work page 2019
[2]

https://www.cvast.tuwien.ac.at/smartct/ sabrina

Sabrina: Modeling and visualization of ﬁnancial data over time with incremental domain knowledge — centre for visual analytics science and technology. https://www.cvast.tuwien.ac.at/smartct/ sabrina. Accessed: 2019-07-25

work page 2019
[3]

https: //www.wu.ac.at/bibliothek/recherche/datenbanken/info/ sabina/

Wirtschaftsuniversitt wien: Sabina - info - datenbanken. https: //www.wu.ac.at/bibliothek/recherche/datenbanken/info/ sabina/. Accessed: 2019-06-10

work page 2019
[4]

http://www.statistik.at/KDBWeb/kdb_ VersionAuswahl.do

nace 2008 - struktur. http://www.statistik.at/KDBWeb/kdb_ VersionAuswahl.do. Accessed: 2019-07-20

work page 2008
[5]

Barrett, C

C. Barrett, C. L. Conway, M. Deters, L. Hadarean, D. Jovanovi’c, T. King, A. Reynolds, and C. Tinelli. CVC4. In G. Gopalakrishnan and S. Qadeer, eds., Proceedings of the 23rd International Conference on Computer Aided Veriﬁcation (CAV ’11), vol. 6806 of Lecture Notes in Computer Science, pp. 171–177. Springer, July 2011. Snowbird, Utah

work page 2011
[6]

Barrett and C

C. Barrett and C. Tinelli. Satisﬁability modulo theories. In Handbook of Model Checking, pp. 305–343. Springer, 2018

work page 2018
[7]

C. P. Birch, S. P. Oom, and J. A. Beecham. Rectangular and hexago- nal grids used for observation, experiment and simulation in ecology. Ecological modelling, 206(3-4):347–359, 2007

work page 2007

Show all 26 references

[8]

Chevaleyre, P

Y . Chevaleyre, P. E. Dunne, U. Endriss, J. Lang, M. Lemaitre, N. Maudet, J. Padget, S. Phelps, J. A. Rodriguez-Aguilar, and P. Sousa. Issues in multiagent resource allocation. Informatica, 30(1), 2006

2006
[9]

Didimo, G

W. Didimo, G. Liotta, and F. Montecchiani. Vis4aui: Visual analysis of banking activity networks. In P. Richard, M. Kraus, R. S. Laramee, and J. Braz, eds., GRAPP/IVAPP, pp. 799–802. SciTePress, 2012

2012
[10]

W. Dong, S. Wang, Y . Chen, and L. Meng. Using eye tracking to evaluate the usability of ﬂow maps. ISPRS International Journal of Geo-Information, 7(7):281, 2018

2018
[11]

Jenny, D

B. Jenny, D. M. Stephen, I. Muehlenhaus, B. E. Marston, R. Sharma, E. Zhang, and H. Jenny. Design principles for origin-destination ﬂow maps. Cartography and Geographic Information Science, 45(1):62–75, 2018

2018
[12]

J. D. Kirkland, T. E. Senator, J. J. Hayden, T. Dybala, H. G. Goldberg, and P. Shyr. The nasd regulation advanced-detection system (ads). AI Magazine, 20(1):55, 1999

1999
[13]

R. A. Leite, T. Gschwandtner, S. Miksch, S. Kriglstein, M. Pohl, E. Gstrein, and J. Kuntner. Eva: Visual analytics to identify fraudulent events. IEEE transactions on visualization and computer graphics , 24(1):330–339, 2018

2018
[14]

Y . Luo, K. Liu, and D. N. Davis. A multi-agent decision support system for stock trading. IEEE network, 16(1):20–27, 2002

2002
[15]

Miksch and W

S. Miksch and W. Aigner. A matter of time: Applying a data–users– tasks design triangle to visual analytics of time-oriented data. Comput- ers & Graphics, 38:286–290, 2014

2014
[16]

T. Munzner. A nested model for visualization design and validation. IEEE transactions on visualization and computer graphics, 15(6):921– 928, 2009

2009
[17]

G. O. Passmore and D. Ignatovich. Formal veriﬁcation of ﬁnancial algorithms. In International Conference on Automated Deduction, pp. 26–41. Springer, 2017

2017
[18]

Pham and S

T. Pham and S. Lee. Anomaly detection in bitcoin network using unsupervised learning methods. arXiv preprint arXiv:1611.03941 , 2016

2016 arXiv
[19]

Poledna, M

S. Poledna, M. Miess, and S. Thurner. Economic forecasting with an agent-based model. 2017

2017
[20]

A. Rae. Flow-data analysis with geographical information systems: a visual approach. Environment and Planning B: Planning and Design, 38(5):776–794, 2011

2011
[21]

J. Rothe. Economics and computation, vol. 4. Springer, 2015

2015
[22]

Sousa, C

P. Sousa, C. Ramos, and J. Neves. The fabricare scheduling prototype suite: Agent interaction and knowledge base. Journal of Intelligent Manufacturing, 14(5):441–455, 2003

2003
[23]

M. P. Timmer, E. Dietzenbacher, B. Los, R. Stehrer, and G. J. De Vries. An illustrated user guide to the world input–output database: the case of global automotive production. Review of International Economics, 23(3):575–605, 2015

2015
[24]

M. P. Timmer, E. Dietzenbacher, B. Los, R. Stehrer, and G. J. de Vries. An illustrated user guide to the world inputoutput database: the case of global automotive production. Review of International Economics, 23(3):575–605, 2015. doi: 10.1111/roie.12178

2015 doi
[26]

Vrotsou, G

K. Vrotsou, G. Fuchs, N. Andrienko, and G. Andrienko. An interactive approach for exploration of ﬂows through direction-based ﬁltering. Journal of Geovisualization and Spatial Analysis, 1(1-2):1, 2017. 5

2017

Pith tools

Reviewed August 14, 2026 · model on record in the stance chip above.

[1] [25]

Tsigkanos, A

C. Tsigkanos, A. Arleo, J. Sorger, and S. Dustdar. How do ﬁrms trans- act? Guesstimation and validation of ﬁnancial transaction networks with satisﬁability. In IEEE 20th International Conference on Informa- tion Reuse and Integration, Los Angeles, California, USA, July 30 - August 1 2019, (to appear)

work page 2019

[2] [1]

https://github.com/keplergl/kepler.gl

kepler.gl. https://github.com/keplergl/kepler.gl. Accessed: 2019-06-12

work page 2019

[3] [2]

https://www.cvast.tuwien.ac.at/smartct/ sabrina

Sabrina: Modeling and visualization of ﬁnancial data over time with incremental domain knowledge — centre for visual analytics science and technology. https://www.cvast.tuwien.ac.at/smartct/ sabrina. Accessed: 2019-07-25

work page 2019

[4] [3]

https: //www.wu.ac.at/bibliothek/recherche/datenbanken/info/ sabina/

Wirtschaftsuniversitt wien: Sabina - info - datenbanken. https: //www.wu.ac.at/bibliothek/recherche/datenbanken/info/ sabina/. Accessed: 2019-06-10

work page 2019

[5] [4]

http://www.statistik.at/KDBWeb/kdb_ VersionAuswahl.do

nace 2008 - struktur. http://www.statistik.at/KDBWeb/kdb_ VersionAuswahl.do. Accessed: 2019-07-20

work page 2008

[6] [5]

Barrett, C

C. Barrett, C. L. Conway, M. Deters, L. Hadarean, D. Jovanovi’c, T. King, A. Reynolds, and C. Tinelli. CVC4. In G. Gopalakrishnan and S. Qadeer, eds., Proceedings of the 23rd International Conference on Computer Aided Veriﬁcation (CAV ’11), vol. 6806 of Lecture Notes in Computer Science, pp. 171–177. Springer, July 2011. Snowbird, Utah

work page 2011

[7] [6]

Barrett and C

C. Barrett and C. Tinelli. Satisﬁability modulo theories. In Handbook of Model Checking, pp. 305–343. Springer, 2018

work page 2018

[8] [7]

C. P. Birch, S. P. Oom, and J. A. Beecham. Rectangular and hexago- nal grids used for observation, experiment and simulation in ecology. Ecological modelling, 206(3-4):347–359, 2007

work page 2007