Open Financial Data from the Macedonian Stock Exchange

Abstract

The concept of Open Data, which represents the idea that public data should be published in a machine-readable format, starts to take a significant role in modern society. Public data from various fields are being transformed in open data formats and published on systems which allow easier consumption from software agents and applications, as well as the users behind them. On the other hand, people in the business world are trying for a few decades now to establishing standards for financial accounting that govern the preparation of financial reports. Financial reporting has crucial significance for companies today, since it is a record of their work which is presented to their stakeholders and represents a starting point for future business decisions and strategies. In this paper, we use data from the Macedonian Stock Exchange and data from different web sites of Macedonian companies in order to create datasets of Open Financial Data relevant for our country, thus increasing the transparency and improving the data accessibility. We describe the process of transforming the data into 4 star Open Data, and present use-case scenarios which use data from our generated datasets and from the World Bank. The datasets are published and accessible via a SPARQL endpoint, and we demonstrate how a software application can make use of them.

Full text

© Springer International Publishing Switzerland 2015
A. Madevska Bogdanova and D. Gjorgjevikj, ICT Innovations 2014,
115
Advances in Intelligent Systems and Computing 311, DOI: 10.1007/978-3-319-09879-1_12
Open Financial Data from the Macedonian Stock
Exchange
Bojan Najdenov, Hristijan Pejchinoski, Kristina Cieva,
Milos Jovanovik, and Dimitar Trajanov
Faculty of Computer Science and Engineering, Ss. Cyril and Methodius University,
Skopje, Republic of Macedonia
{bojan.najdenov,milos.jovanovik,dimitar.trajanov}@finki.ukim.mk,
hristijan.pejcinoski@gmail.com, kristina_cieva@yahoo.com
Abstract. The concept of Open Data, which represents the idea that public data
should be published in a machine-readable format, starts to take a significant
role in modern society. Public data from various fields are being transformed in
open data formats and published on systems which allow easier consumption
from software agents and applications, as well as the users behind them. On the
other hand, people in the business world are trying for a few decades now to
establishing standards for financial accounting that govern the preparation of
financial reports. Financial reporting has crucial significance for companies
today, since it is a record of their work which is presented to their stakeholders
and represents a starting point for future business decisions and strategies. In
this paper, we use data from the Macedonian Stock Exchange and data from
different web sites of Macedonian companies in order to create datasets of
Open Financial Data relevant for our country, thus increasing the transparency
and improving the data accessibility. We describe the process of transforming
the data into 4 star Open Data, and present use-case scenarios which use data
from our generated datasets and from the World Bank. The datasets are
published and accessible via a SPARQL endpoint, and we demonstrate how a
software application can make use of them.
Keywords: Finances, Open Data, Macedonian Stock Exchange, World Bank,
RDF, Ontologies.
1 Introduction
The main idea that lies behind the concept of Open Data1 is that public data should be
free and available to everyone. We live in a world where information holds great
value. Having the right information at the right time, in the right way, builds modern
societies, drives technologies forward, develops businesses and even saves lives. The
exponential growth of datasets about people, technological artifacts and organizations
brought us in position where we have on disposal vast amounts of information ready
1 http://okfn.org/opendata/

116 B. Najdenov et al.
to be rearranged and shap
e
the structured, machine-re
a
becoming the future of star
t
Linked Open Data
2
is a
sufficiently interlinked dat
a
of large-scale datasets hav
e
constantly [2]. As we see i
n
in Linked Open Data for
m
community [3].
The concept of Linked
stored on different locatio
n
OWL and SPARQL. By
u
data silos can be successfu
l
barriers which occur durin
g
data analysis by simplifyi
n
from various industry field
e.g. regulatory bodies and
b
Financial accounting is
p
the companies’ work. Th
e
amount of datasets. Thus,
t
for representation, storage,
big importance for today’s
a practical solution usin
g
Macedonian Stock Excha
n
publish on their websites.
Fig.
1
2
http://linkeddata.
o
3
http://lod-cloud.n
e
e
d in order to create additional value [1]. This implies
t
a
dable, open data that is free to access and interlin
k
t
up companies and business in general.
community effort to alleviate the problem of the lac
k
a
sets on the Web. Through this effort, a significant nu
m
e
now been published in the LOD cloud
3
, which is gro
w
n
Fig. 1, datasets from different fields are publicly avail
a
m
at, thanks to the contributors to the Linked Open
D
Open Data provides us with a way to connect data
n
s, by using the Semantic Web standards such as R
D
u
sing the existing Web infrastructure, data from diffe
r
l
ly interconnected and the Web can be used to decrease
g
process of linking [4]. Linked Open Data enables b
e
n
g the process of combining information sources. Data
s can then be used in different ways and by many enti
t
b
anks, when it comes to financial data [5].
p
rimarily oriented towards creating the financial reports
e
process of creating the reports is dependent on a h
t
his is a field that necessarily requires different approa
c
querying and visualizing of the data. We find this issu
e
companies and economies which motivated us to wor
k
g
data about Macedonian companies provided by
n
ge, the World Bank and the information and data
t
1
. The LOD Cloud, as of September 2011
o
rg/
e
t/
t
hat
k
, is
k
of
m
ber
w
ing
a
ble
D
ata
sets
D
F,
r
ent
the
e
tter
sets
t
ies,
for
h
uge
c
hes
e
of
k
on
the
t
hey

Open Financial Data from the Macedonian Stock Exchange 117
2 Related Work
Numerous projects exist which have a major target to either publish financial or
corporate data in Open Data formats, or enable their annotation with the technologies
of the Semantic Web, in order to leverage their value. The World Bank, as one of the
most important financial institutions on a global level, puts great effort in many
projects which result with creating Open Data. Other significant projects in this area
are the Financial Industry Business Ontology (FIBO), the Open Corporates project
and the Financial Report Ontology.
The World Bank aims towards decreasing extreme poverty in the world, through
proving financial and technical assistance to developing countries. The financial support
the developing countries receive is in form of low-interest loans, credits and grants, or
investments in various areas like healthcare, education, infrastructure, resource
management etc. The World Bank, as a global institution, supports the ideas behind the
Open Data concept, and therefore shares its public data freely on their website4.
In [6], the authors introduce an interesting project which aims towards designing
new methods for extraction of data and, based on that, developing a prototype for
extracting financial information from the semi-structured text. They believe that in the
financial world numbers are often one main target, but they are meaningless without
any semantic meta-data describing what kind of information they represent.
The Financial Industry Business Ontology5 (FIBO) is an initiative to define and
describe terms and rules for financial data. Its goal is building a representation of the
information about financial instruments, market data, business entities, etc. along with
the relationships between them.
Open Corporates6 is one of the largest open databases of companies in the world,
having information about 63 million companies from around the globe. They publish
the data in XML, RDF or JSON format and it can be downloaded from their website.
They believe that basic corporate information about all the companies in the world
should be brought together in one place, making it easier to access, use and connect
with other data.
The Financial Report Ontology7 is a project developed with the idea of providing
an ontology that would describe the financial reports as concepts, as well as their
individual entries. The ontology aims to assist companies in the process of creating
annotated financial reports.
3 Macedonian Open Financial Data
3.1 Public Data from the Macedonian Stock Exchange
The Macedonian Stock Exchange (MSE)8 is the only financial institution in
Macedonia that is authorized to organize, execute and regulate the trading of
4 http://data.worldbank.org/
5 http://www.omg.org/hot-topics/fibo.htm
6 http://opencorporates.com/
7 http://financialreportontology.wikispaces.com/
8 http://www.mse.mk/en/

118 B. Najdenov et al.
securities. It was established in 1995 as a joint stock company and the first trading
occurred in March, 1996. The main purpose of MSE is to provide security and
efficiency in the organized trading of securities in Macedonia.
MSE is comprised of two market segments: Official Market and Regular Market.
The stock market indices are MBI10 (Macedonian Blue Chip Index), which includes
the stocks of the 10 most traded companies, MBID (Macedonian Stock Exchange
Index of publicly held companies), which includes the stocks of the publicly held
companies listed on MSE and OMB (Bond Index), which includes issued bonds listed
on MSE.
MSE publishes most of its data on their website, either as PDF files or in HTML
tables. Among all of the published data, like stock prices, different indices,
information about growth trends on securities, etc., our main topic of interest are the
financial reports which MSE member companies publish. We gathered the financial
report data from the MSE website, converted it and stored it in CSV format. We did
the same process for gathering and storing the company data, which we obtained from
individual companies websites.
3.2 Open Data from the World Bank
As we already noted, the World Bank published data from its projects on their
website. Parts of these data are the financial data, which allow us to see what global
funds the World Bank manages, visualize them or build models over them.
Many different financial datasets can be found on World Bank’s website9 in
various different formats, such as CSV, JSON, PDF, RDF, RSS, XLS, XLSX and
XML. Some of their datasets can be accessed via the public SPARQL endpoint which
the World Bank provides10, as part of their Linked Data project. The dataset that we
are interested in contains data on commitments against contracts that were reviewed
by the Bank before they were awarded (prior-reviewed Bank-funded contracts) under
IDA/IBRD11 investment projects and related Trust Funds. We downloaded this
dataset in RDF format and linked its data with data published by MSE and
Macedonian companies. The procedure will be described in details.
4 Ontologies for the Datasets
4.1 Ontology for the World Bank Dataset
We loaded the dataset from the World Bank data store into a local Virtuoso Universal
Server12 instance, as an RDF graph. Since all the entries in the dataset refer to a loan
awarded to a company by the World Bank, a single entry in the dataset can be
considered as a resource which provides all the details related to a specific loan.
9 https://finances.worldbank.org/all-datasets
10 http://worldbank.270a.info/sparql
11 http://data.worldbank.org/indicator/DT.DOD.MWBG.CD
12 http://virtuoso.openlinksw.com/

Open Financial Data from the Macedonian Stock Exchange 119
4.2 Corporate Registry Ontologies
As we already mentioned, Open Corporates holds a large publicly available dataset of
information about companies as legal entities, for all around the world. Unfortunately,
they do not hold any information about Macedonian companies, and therefore we
cannot use their datasets in the context of Macedonian financial data.
However, we did analyze their data and the ontologies they use for semantic
annotation, so we decided to reuse those ontologies and annotate our data in a similar
manner. Another motivation for this was the similarity between the structures of the
dataset from Open Corporate had with the data we were able to collect for
Macedonian companies. The ontologies we use in describing the companies as legal
entities are listed in Table 1.
Table 1. The ontologies we reused for Macedonian company data
Prefix URI
foaf http://xmlns.com/foaf/0.1/
vCard http://www.w3.org/2006/vcard/ns#
adms http://www.w3.org/ns/adms#
rov http://www.w3.org/ns/regorg#
skos http://www.w3.org/2004/02/skos/core#
We use the rov:RegisteredOrganization class in order to represent a legal entity or
organization which is legally registered, i.e. a company that we have data about. The
rest of the DataType properties we use to describe a Registered Organization can be
found in Table 2.
Table 2. The DataType properties we use
Property Description
rov:legalName The legal name of the company.
rov:registration The registration is a fundamental relationship between a legal
entity and the authority with which it is registered and that
confers legal status upon it. rov:registration is a sub property
of adms:identifier which has a range of adms:Identifier.
vCard:extended-address The address of the object.
vCard:hasTelephone To specify the telephone number for telephony
communication with the object.
skos:notation Refined name of a company.
foaf:homepage A homepage for some company. Every value of this property
is a foaf:Document.
rdfs:label Information about the basic activities of a company.

120 B. Najdenov et al.
4.3 Financial Report Ontology
Every member of the Macedonian Stock Exchange provides annual financial reports
which are the balance sheet, income statement, statement of cash flows and the
statement of retained earnings. Our focus in this paper is the balance sheet of the
companies in particular, which requires an ontology to be provided so that we could
semantically annotate that data.
For this purpose we decided to reuse the Financial Report Ontology which, as we
already described, defines the basic financial report terms.
In the ontology we find the class Fundamental Accounting Concept, which
represents one full financial report. Its properties are divided into five groups: General
Information properties, Balance Sheet, Income Statement, Statement of
Comprehensive Income and Cash Flow Statement properties. For our local reports we
will use only General Information, Balance Sheet and Income Statement properties.
Table 3. The properties in the CFRL ontology
Property Description
cfrl:hasReport This property connects a company i.e. instance of
RegisteredOrganization class, with its financial report.
cfrl:hasLoan This property points to the World Bank loans that are made by
that company.
4.4 Corporate Financial Reports and Loans Ontology
In order to be able to successfully complete the annotation and linking process
between the datasets, we developed the Corporate Financial Reports and Loans
Ontology (CFRL). In it, we introduce two object properties: “hasReport” and
“hasLoan”. Their main role is to provide means of interlinking the datasets. The
description of these two properties can be found in Table 3.
5 Linking the Datasets
Before we begin explaining the process of interlinking the datasets, we must state that
our goal is to interlink the data from our corporate registry dataset, i.e. the data we
gathered from various websites of different companies, with the data we acquired
from the World Bank about loans that companies were awarded and also with the
financial reports data we got from the Macedonian Stock Exchange. Conceptually, the
linking we wish to achieve is shown in Fig. 2.

Op
e
5.1 Mapping the Data
f
The next step of our work
RDF and to do that, w
e
mechanisms for data transf
o
The technical process o
f
done using the R2RML
m
registry dataset and the fin
a
details originating from th
e
it directly in the Virtuoso
U
5.2 Interlinking the R
D
Having transformed all the
interlink the data, as show
n
our CFRL ontology: “cfrl:
h
The “cfrl:hasReport” pr
o
That means, we connect a
“
“FundamentalAccountingC
o
this purpose we use the “s
k
and the “fac:EntityRegistr
a
entity.
The property “cfrl:hasL
o
loan entities. Similarly to t
h
the names of the corr
e
“rov:legalName” of an “R
e
property of a loan entry.
These interlinking proce
The resulting linked d
a
SPARQL endpoint
14
.
13
http://www.w3.org/
T
14
http://linkeddata.
f
e
n Financial Data from the Macedonian Stock Exchange
Fig. 2. Linking the datasets
f
rom CSV to RDF
is mapping and transforming datasets from CSV file
s
e
use the Virtuoso Universal Server, which prov
i
o
rmation management and querying using SPARQL.
f
mapping and transforming the data from CSV to RDF
w
m
apping language
13
, as described in [7], for the corpo
r
a
ncial reports dataset, respectively. The dataset about l
e
World Bank was already in RDF format, so we impo
r
U
niversal Server instance, as an RDF graph.
D
F Datasets
datasets into RDF graphs in Virtuoso, our next step w
a
n
in Fig. 2. For that purpose of we created two properti
e
h
asReport” and “cfrl:hasLoan”.
o
perty links a company with its published financial rep
o
“
RegisteredOgranization” entity with its financial report
o
ncep
t
”, by matching the values of the company name.
k
os:notation” property of a “RegisteredOgranization” en
t
a
ntName” property of a “FundamentalAccountingConc
e
o
an” interlinks a “RegisteredOgranization” entity wit
h
h
e previous property, we create the connection by matc
h
e
spondent companies. For this purpose we use
e
gisteredOrganization” entity, and the “worldbank:suppl
sses were done using SPARQL queries over the dataset
a
ta that we generated, can be accessed through a pu
b
T
R/r2rml/
f
inki.ukim.mk/sparql
121
s
to
i
des
w
as
r
ate
oan
r
ted
a
s to
e
s in
o
rts.
i.e.
For
t
ity,
e
p
t
”
h
its
h
ing
the
l
ier”
s.
b
lic

122 B. Najdenov et al.
6 Use-Cases
The main purpose of using interlinked Open Data datasets is the ability to increase the
value and usability of the separate datasets, by providing advances use-case scenarios.
We are going to describe two of the many possible scenarios.
6.1 Displaying Information from the World Bank
We demonstrate the use of the “hasLoan” property to retrieve information about a
company which obtained a loan from the World Bank, or to be more precise, the dates
when the company signed contracts for getting loans with the World Bank, the total
contract amount (USD) and which sector was the loan dedicated to. For the purpose
of the demonstration, we show the top 5 loans and their details. The SPARQL query
is the following:
prefix cfrl: <http://linkeddata.finki.ukim.mk/lod/ontology/cfrl#>
prefix worldbank: <http://finances.worldbank.org/resource/>
prefix rov: <http://www.w3.org/ns/regorg#>
SELECT ?s ?csd ?tca ?ms WHERE {
?company rov:legalName ?s .
?s cfrl:hasLoan ?l .
?l worldbank:contract_signing_date ?csd ;
worldbank:supplier_contract_amount_usd ?tca ;
worlbank:major_sector ?ms .
} ORDER BY DESC (?tca) LIMIT 5
The result of the executed query at our Virtuoso SPARQL endpoint, are shown in
Table 4.
Table 4. Results from the SPARQL query
Supplier Contract signing date Total contract amount Major sector
Granit Mar 26, 2009 $9,802,524.00 Transportation
Granit Dec 04, 2009 $6,197,108.00 Transportation
Granit Dec 04,2009 $5,323,028.00 Transportation
Granit Mar 26, 2009 $4,519,095.00 Transportation
Granit Dec 04,2009 $3,785,761.00 Transportation
6.2 Displaying Information from the Financial Reports
In this section we show how the “hasReport” property that we defined in our CFRL
ontology, can be used to provide additional information about companies. One such
scenario would be to retrieve information about the top 5 companies by the profit they

Open Financial Data from the Macedonian Stock Exchange 123
made in the year of 2012, in Macedonian Denars (MKD). For that purpose we can use
the following SPARQL query:
prefix cfrl: <http://linkeddata.finki.ukim.mk/lod/ontology/cfrl#>
prefix fac:
<http://www.xbrlsite.com/2013/FinancialReportOntology/Prototype04/FundamentalAc
countingConcepts.xml#>
prefix rov: <http://www.w3.org/ns/regorg#/>
SELECT ?name ?profit ?period WHERE {
?cmp cfrl:hasReport ?rep ; rov:legalName ?name .
?rep fac:GrossProfit ?profit ; fac:FiscalPeriod ?period .
FILTER (?period = 2012)
} ORDER BY ?profit LIMIT 5
The result of this query, showing the name of such companies and the profit they
made in the year of 2012, can be seen in Table 5.
Table 5. Results from the SPARQL query
Name Profit (MKD) Period
ALKALOID AD SKOPJE 3,291,423 2012
Stopanska Banka AD Skopje 2,376,477 2012
Tikvesh AD Skopje 339,049 2012
GD GRANIT AD - Skopje 291,238 2012
Vitaminka AD Prilep 102,378 2012
7 Conclusion and Future Work
Data, information and knowledge management are key activities in modern
economies and considerable efforts and resources are devoted for research in these
areas, by different organizations in the world. Having data structured and interlinked
provides a whole new area of opportunities for data usage and management. This
provides huge benefits in the information dissemination processes and provides
mechanisms so that information can be shared easily between bank divisions,
institutions and distributed to all stakeholders.
In this paper we gave an overview of the process of transforming the one-star and
two-star data about companies into four-star Open Data and connected it with a
dataset from the World Bank. We also provided use-case scenarios which gave
examples of how our local data and how the data from the World Bank can be used in
order to provide information which is not available when the datasets are isolated.
With this, we hope our work contributes to the goals of the Open Data Initiative15 in
Macedonia.
15 http://opendata.gov.mk/

124 B. Najdenov et al.
In the future, we plan to continue our work in these fields, increase the amount of
datasets, connect our data with other remote resources and transform these datasets
further to five-star data, interlinked with financial data published on the LOD cloud.
This would improve the quality of the use-cases we provide and also create new
opportunities for development of creative applications and analysis. We hope our
work serves as a motivation to companies, financial institutions, organizations around
the world, to recognize the benefits of open financial data and publish their public
data on the Web in raw and machine-readable format.
Acknowledgment. The work in this paper was partially financed by the Faculty of
Computer Science and Engineering, at the Ss. Cyril and Methodius University in
Skopje, as part of the research project “Semantic Sky 2.0: Enterprise Knowledge
Management”.
References
1. Cardoso, J., Pedrinaci, C., Leidig, T., Rupino, P., De Leenheer, P.: Open semantic service
networks. In: International Symposium on Services Science (ISSS), Leipzig, Germany
(2012)
2. Möller, K., Hausenblas, M., Cyganiak, R., Handschuh, S., Grimnes, G.: Learning from
Linked Open Data Usage: Patterns & Metrics. In: Web Science Conference (WSC) (2010)
3. Bizer, C., Heath, T., Berners-Lee, T.: Linked Data - The Story So Far. International Journal
on Semantic Web and Information Systems (IJSWIS), 1–22 (2009)
4. Kundra, V.: Digital Fuel of the 21st Century: Innovation through Open Data and the
Network Effect. Joan Shorenstein Center on the Press, Politics and Public Policy, Harvard
College (2012)
5. Radzimski, M., Sánchez-Cervantes, J.L., Rodríguez-González, A., Gómez-Berbís, J.M.,
García-Crespo, A.: FLORA –Publishing Unstructured Financial Information in the Linked
Open Data Cloud. In: First International Workshop on Finance and Economics on the
Semantic Web (FEOSW) (2012)
6. Bjoraa, E.: Ontology guided financial knowledge extraction from semi-structured
information sources. Master Thesis in Information and Communication Technology, Agder
University Colledge, Grimstad (May 2003)
7. Jovanovik, M., Najdenov, B., Trajanov, D.: Linked Open Drug Data from the Health
Insurance Fund of Macedonia. In: 10th International Conference for Informatics and
Information Technology (2013)