Applying Text Analytics Methodology to Analyze Project Reports

Irina Arhipova

, Liga Paura

, Nikolajs Bumanis

, Gatis Vitols

, Vladimirs Salajevs

Aldis Erglis

, Gundars Berzins

and Evija Ansonska

Faculty of Engineering and Information Technologies, Latvia University of Life Sciences and Technologies,

Liela Street 2, Jelgava, LV 3001, Latvia

Faculty of Business, Management and Economics, University of Latvia, Aspazijas bulv. 5, Riga, LV 1050, Latvia

{aldis.erglis, gundars.berzins, evija.ansonska}@lu.lv, aldis.erglis@outlook.com

Keywords: Text Mining, Text Gap Analysis, Word Co-Occurrence Analysis, Unique Terms Identification.

Abstract: The goal of this article is to develop a support methodology for Driving Urban Transition (DUT) partnership

to ensure that the knowledge gathered from ERA-NET Urban Accessibility and Connectivity (EN-UAC,

2023) projects, to repeatedly identify the requirements of programme entities and define specific topics for

future calls. Fifteen projects under the Horizon 2020 ERA-NET initiative have been analysed to detect

similarity between projects, uniqueness of the projects, project compliance with DUT and SRIA, and gap

between projects and DUT, SRIA methodology. A particular focus in the analysis was on the project

“Individual Mobility Budgets as a Foundation for Social and Ethical Carbon Reduction” (MyFairShare). Text

mining methods were used for documents analysis. The similarity between the documents detected by the

cluster algorithm and they were compared using words, as a result, the documents were combined into three

clusters: “Strategy implementation and network infrastructure”; “Transport accessibility and policy” and

“Urban city mobility”. The identification of unique terms shown the terms energy, ecosystem and climate are

unique for DUT&SRIA and are not found in 15 EN-UAC project applications and the next specific topics for

future calls can be within the subject of energy, climate and ecosystem.

1 INTRODUCTION

The objective of numerous European, national,

regional, and local transportation strategies and

measures is to offer economical, reachable, secure,

and dependable urban access and linkage while

imposing minimal or no harm on the environment.

Over time, the European Commission, along with

member states, regional and local governing bodies,

private enterprises, and other parties involved, have

diligently worked to propel the progression of

mobility and transportation systems. Nevertheless,

despite these endeavours, sustainability predicaments

persist, underscoring the immediate requirement for a

profound shift toward a carbon-neutral, equitable, and

https://orcid.org/0000-0003-1036-2024

https://orcid.org/0000-0002-6625-9475

https://orcid.org/0000-0002-4131-8635

https://orcid.org/0000-0001-8545-4690

https://orcid.org/0000-0002-1302-527X

https://orcid.org/0000-0002-7029-2745

emission-free transportation network to climate-

neutral and smart cities (Clerici Maestosi et al., 2021).

The 15-Minute City (15minC) (Moreno, 2019),

Downsizing District Doughnuts (DDD) and Positive

Energy Districts (PED) are three innovation pillars or

new approaches for urban greening (Clerici Maestosi

et al., 2021). In 121 European metropolitan areas the

15-minute walking city (15-MWC) index was

evaluated and authors concluded the cities in the UK,

Portugal, France, and Italy have a lower 15-MWC

index and worst accessibility to their daily needs

compared to West Germany (Bartzokas-Tsiompras &

Bakogiannis, 2023). Along with the research on the

15minC, the researches on walking accessibility

(Bartzokas-Tsiompras et al., 2023; Ione Avila-

Palencia et al., 2022) and the behaviour of the

Arhipova, I., Paura, L., Bumanis, N., Vitols, G., Salajevs, V., Erglis, A., Berzins, G. and Ansonska, E.

Applying Text Analytics Methodology to Analyze Project Reports.

DOI: 10.5220/0012555100003717

Paper published under CC license (CC BY-NC-ND 4.0)

In Proceedings of the 6th International Conference on Finance, Economics, Management and IT Business (FEMIB 2024), pages 89-95

ISBN: 978-989-758-695-8; ISSN: 2184-5891

inhabitants (Arhipova et al., 2023) remains even more

relevant.

The urgent and essential transformation of our

planet demands an immediate and critical shift in

society. This shift necessitates individuals to re-

evaluate their behaviours and outlooks, which in turn

involves a process of learning. Within the framework

of transitioning toward entirely new socio-economic

practices as envisioned by sustainable urban mobility,

learning goes beyond the mere acquisition of

knowledge.

Establishing an environment that encourages the

exploration of diverse forms of knowledge and

various ways of understanding is vital to cultivate

learning centred on sustainability. This learning

revolves around the notion of accessibility and

connectivity (ACUTE, 2023).

The goal of this article is to develop support

methodology for Driving Urban Transition (DUT)

partnership to ensure that the knowledge gathered

from ERA-NET Urban Accessibility and

Connectivity (EN-UAC, 2023) projects and beyond is

accessible as assistance for all stakeholders within the

forthcoming Horizon Europe Partnership Program

(DUT, 2023). This support aims to drive the urban

transition toward a future that is both sustainable and

conducive to a high quality of urban living.

In order to fulfil the goal of the article, it is

essential to repeatedly identify the requirements of

programme entities and define specific topics. This

involves addressing research gaps, implementation

deficiencies, using existing outcomes as a foundation

for future calls, staying abreast of new developments

and trends in urban accessibility and connectivity

innovations and social practices, and exploring

potential synergies with other funding programmes.

2 METHODOLOGY

Usually the text is viewed as a means of conveying

meaning, but it can also be seen as a structured

sequence of words or an unstructured collection of

words. These words can be represented in ways that

enable analysis without being limited by grammatical

structure. Texts can be subjected to statistical

examination by exploring the interrelationships

between words within a specific text and comparing

these relationships to those observed in other texts

(Robinson & Silge, 2017). The utilization of

computational methods involves:

 complementary enhancement to qualitative

analysis approaches;

 scalability, capable of processing numerous

documents (such as proposals, mid-term

reports, final reports, etc.);

 application of text analysis method: unordered

bag of words;

 employment of data mining techniques to

unearth novel insights for upcoming calls:

ordered string of words;

 manual customization options, allowing for the

definition of stop-words and precision;

 while not flawless, it operates in a semi-

automated manner.

The set (=corpus) of 15 ERA-NET Urban

Accessibility and Connectivity (EN-UAC) project’

proposals (Basecamp, 2023), Driving Urban

Transitions to a Sustainable Future Roadmap 2022

(DUT, 2023) and JPI Urban Europe’s Strategic

Research and Innovation Agenda texts (=documents)

is analysed (SRIA 2.0., 2023).

The process of text analytics or text mining

involves multiple sequential steps (see Figure 1). The

initial step is text normalization by eliminating

punctuation, cases, and numbers. It is essential to

remove common words, often referred to as

stopwords, as they can obscure the more significant

and relevant ones. This includes removing standard

English stopwords (Vijayarani et al., 2015).

The next step is words frequency analysis: data

visualization by generating Wordclouds. After that

follows text similarity analysis: the idea is to measure

the similarity between two words or texts in order to

cluster documents into distinct groups. The ultimate

goal is to pinpoint a specific collection of documents

that maximizing the level of similarity. A similarity

model can be employed not only for clustering

documents but also for clustering words.

Word co-occurrence analysis step is necessary to

check co-occurrence between DUT/SRIA and

projects' proposals the combinations of files (entities

in corpus) are created, for example, DUT –

MyFairShare project proposal text, SRIA 2.0 –

MyFairShare project proposal text (Arhipova et al.,

2022), etc. and for each combination:

 a corpus, containing two files is created;

 out of each file, 10 most used terms are

selected;

 the list of unique words is created, using 10

most used terms from each file;

 unique term combinations are created using

unique words;

 each combination is then checked for

belonging to corpus entities;

FEMIB 2024 - 6th International Conference on Finance, Economics, Management and IT Business

 the results are stored as matrix where

X=Y=unique word and XY value equals to the

number of occurrences of particular

combination.

As the result of word co-occurrence analysis, the

matrix, as .csv file, is passed as input file to

visualization software Gephi (Gephi, 2023).

The last step of the process of text analytics is

unique terms identification and text gap analysis. The

word frequency provides representations of content

that include commonly occurring terms. However, to

identify topics that are specific to each text, it is

necessary a different approach.

The most commonly employed method for

achieving this is by utilizing the term frequency-

inverse document frequency (TF-IDF) measure,

which stands for term frequency-inverse document

frequency (Robinson & Silge, 2017).

Word frequency, text similarity analysis, unique

terms identification, and text gap analysis was done

using R (R Core Team, 2021) and R Text Mining

Package tm (Feinerer, Hornik, 2020).

Figure 1: The text analytics scheme.

3 RESULTS

Fifteen projects under the Horizon 2020 ERA-NET

Cofund Urban Accessibility and Connectivity (EN-

UAC) initiative have been analysed, with particular

focus on the MyFairShare project titled “Individual

Mobility Budgets as a Foundation for Social and

Ethical Carbon Reduction” (MyFairShare, 2023).

The research devised a theoretical notion of

“equitable” mobility budgets, considering an

individual's social and spatial circumstances (such as

socioeconomic status and proximity to places of

interest). These budgets assist individuals in

comprehending the consequences of their mobility

decisions, and also provide policymakers with

insights into necessary interventions to alleviate

excessive strain on local mobility budgets (Arhipova

et al., 2023) (for example, by enhancing carbon-

neutral accessibility to various destinations).

3.1 Word Frequency and Association

Analysis

Word frequency analysis is done by visualizing data

through Wordcloud generation for EN-UAC projects

proposals, Driving Urban Transitions (DUT, 2023) to

a Sustainable Future Roadmap 2022 and JPI Urban

Europe’s Strategic Research and Innovation Agenda

(SRIA 2.0., 2023) texts (see Figure 4).

The most commonly recurring 20 words (based on

relative frequency) within the texts of 15 EN-UAC

project proposals, DUT, and SRIA 2.0 were

extracted. An instance of this can be observed in the

MyFair Share project, outlined as follows: mobility

(4.9%), budget (3.8%), different (1.6%), policy

(1.5%), measure (1.1%), climate (1.0%), fair (1.0%),

impact (1.0%), lab (1.0%), living (1.0%), transport

(1.0%), etc. (see Figure 2)

Figure 2: Created word cloud of EN-UAC project

MyFairShare.

Correlation was used for the word association

analysis to evaluate six more frequent words in

MyFairShare project (mobility, budget, policy,

climate and fair) around the prominent themes.

The output had shown that mobility and budget

occur in 38% and 56% of the time with the word

calculator, respectively. The most frequent word

association with correlation 0.56 is between words

mobility and budget. The word climate correlated

with words awareness (0.62), protection (0.45) and

change (0.33). The fair correlated with words

distribution (0.37), specific (0.36) and determining

(0.36). This indicates the project main content can be

to determinate mobility budget and it fair distribution.

Applying Text Analytics Methodology to Analyze Project Reports

Figure 3: Created wordcloud of DUT document.

The most frequently five words for DUT are urban

(4.2%), the Positive Energy Districts Transition

Pathway or ped (2.0%), city (2.0%), energy (1.5%)

and innovation (1.4%) (see Figure 3).

Figure 4: Created wordcloud of SRIA 2.0. document.

The most frequently five words for SRIA 2.0. are

urban (8.0%), innovation (2.0%), city (1.5%), agenda

(1.4%), and strategic (1.4%) (see Figure 4).

3.2 Similarity Analysis

The goal of the text similarity analysis is to measure

the similarity between words or texts, enabling the

grouping of documents into distinct clusters. The

objective is to identify a specific set of documents that

maximise similarity levels. In order to measure the

similarity between texts, the hierarchical cluster

algorithm is used for clustering words and

documents. Ward’s or minimum variance method

was used to analyse distance between clusters, which

is popular in linguistic field (Szmrecsanyi, 2012).

The hierarchical clustering results by project

proposals text is shown by dendrogram (see Figure

5).

Figure 5: Dendrogram of 15 EN-UAC projects proposal,

DUT and SRIA 2.0. texts.

Figure 6: Dendrogram of 15 EN-UAC projects proposal,

DUT and SRIA 2.0. words.

Based on the comparison with DUT and SRIA

2.0, it can be concluded that the projects exhibiting

the highest similarity are as follows: CATAPULT and

SmartHub, then Ex-TRA, ASAP, DyMoN,

GeoSence, MyFairShare, TuneOurBlock,

COCOMO, EASIER, WalkUrban, TAP,

SortedMobility, ITEM, JUSTICE, SRIA and DUT

(see Figure 4).

Similarity models can be used not only for

document clustering but also for word clustering. The

clustering result by EN-UAC project proposals, DUT

and SRIA 2.0 words is shown by dendrogram (see

Figure 6).

According to the clustering result the projects

proposals, DUT and SRIA 2.0 words are divided to

the tree clusters. Terms grouped near to each other are

more frequently found together. It can be inferred that

words sharing the greatest similarity can be grouped

together into the following clusters of similarity:

 1

cluster: approach, system, impact, base, tool,

challenge, case, potential, support,

development, service, design, include,

knowledge, stakeholders, solution, work,

sustainable, exit, provide;

FEMIB 2024 - 6th International Conference on Finance, Economics, Management and IT Business

 2

cluster: public, transport, accessibility,

policy;

 3

cluster: mobility, city, urban (see Figure 6).

The names of the clusters’ can be: 1st cluster

“Strategy implementation and Network

infrastructure”; 2nd cluster “Transport accessibility

and policy” and 3rd cluster “Urban city mobility”.

Based on the words analysis between DUT and

SRIA documents three clusters in total with ten most

frequent words was found (see Figure 7).

Word Urban is higher in the plot compere to other

words that appear urban are more common within the

corpus. Word PED (Positive Energy Districts

Transition Pathway) is only occurs in DUT document

and as well as word energy is more frequent in DUT.

Figure 7: Dendrogram of DUT and SRIA 2.0. words.

The results show that the most common words for

DUT and SRIA are urban, city, innovation, energy &

transition, but for ENUAC projects - mobility & city.

The compliance or main subjects for

DUT&SRIA&EN-UAC projects are Strategy and

model development, implementation & network

infrastructure; Transport accessibility & policy;

Urban city mobility.

3.3 Word Co-Occurrence Analysis

The word co-occurrence analysis (Lancia, 2007)

between DUT, SRIA 2.0 and 15 projects' proposals

are executed. Co-occurrence analysis is focused on to

find out words relationship network and words, which

are frequently used together. Force Atlas algorithm is

used for graphs’ layout calculation.

The weight of each connection is the visually

differentiated: the stronger the line, the stronger the

connection (i.e., more occurrences), while the

position of the term depends on its

authority/importance.

Figure 8: Compliance of MyFairShare project proposal to

DUT.

Figure 9: Compliance of MyFairShare project proposal to

SRIA 2.0.

Compliance and gap of MyFairShare project

proposal, DUT and SRIA 2.0., using word co-

occurrence analysis, is given in Figure 8 and Figure

3.4 Unique Terms Identification and

Text Gap Analysis

While word frequency provides representations of

content that include commonly used terms, the

identification of topics specific to each text is

achieved using the TF-IDF measure, which stands for

term frequency-inverse document frequency.

For topics identification that are specific to each text

the TF-IDF measure is used and is calculated as TF

and IDF product:

Applying Text Analytics Methodology to Analyze Project Reports

TF-IDF = TF * IDF,

where

(1)

𝐼𝐷𝐹 

log 

𝑁𝑜. 𝑜𝑓 𝑑𝑜𝑐𝑢𝑚𝑒𝑛𝑡𝑠 𝑖𝑛 𝑐𝑜𝑟𝑝𝑢𝑠

𝑁𝑜. 𝑜𝑓 𝑑𝑜𝑐𝑢𝑚𝑒𝑛𝑡𝑠 𝑖𝑛 𝑐𝑜𝑟𝑝𝑢𝑠 𝑐𝑜𝑛𝑡𝑒𝑖𝑛 𝑡𝑒𝑟𝑚𝑠



(2)

𝑇𝐹 

𝑁𝑜. 𝑜𝑓 𝑡𝑖𝑚𝑒𝑠 𝑡𝑒𝑟𝑚𝑠 𝑎𝑝𝑝𝑒𝑎𝑟𝑠 𝑖𝑛 𝑑𝑜𝑐𝑢𝑚𝑒𝑛𝑡

𝑡𝑜𝑡𝑎𝑙 𝑁𝑜. 𝑜𝑓 𝑡𝑒𝑟𝑚𝑠 𝑖𝑛 𝑑𝑜𝑐𝑢𝑚𝑒𝑛𝑡

(3)

Based on the analysis, the gap between DUT,

SRIA, and 15 EN-UAC projects can be observed in

the following 10 words: energy, regenerative,

ecosystem, neutral, climate, transformation,

economy, dilemma, transition and joint (Figure 10).

The highest TF-IDF (tf-idf) values

for

DUT &

SRIA documents were PED, energy, CURE, climate

and 15minC (Figure 10). Comparison with DUT

Roadmap text & word co-occurrence analysis shows

that energy & transition are words with a lower co-

occurrence of two adjacent terms in a text corpus. The

similar result is found in the identification of unique

terms for DUT&SRIA: energy, ecosystem & climate.

The gap between DUT&SRIA and ENUAC projects

lies within the subject of energy, climate and

ecosystem.

Figure 10: TF-IDF of unique terms for DUT & SRIA

documents.

Figure 11: TF-IDF of unique terms for MyFairShare

proposal.

The top 10 unique words for MyFairShare project

are budget, fair, living lab, protection, climate,

scenario, guidance, distribution and constraint

(Figure 11).

4 CONCLUSIONS

Published research suggest that text mining methods

significantly reduced manual work compared to

conventional methods as well as also enables us to

interpret initial information of the projects proposals,

to estimate bias of the documents and to understand

the connections between them.

Our analysis indicates the most common words

for DUT and SRIA are urban, city, innovation, energy

and transition and for EN-UAC projects are mobility

and city. Based on text analysis the compliance or

main subjects for EN-UAC projects are strategy and

model development, implementation and network

infrastructure; transport accessibility and policy;

urban city mobility.

The identification of unique terms for

DUT&SRIA shows that the terms energy, ecosystem,

and climate are not found in 15 EN-UAC project

applications. For MyFairShare project words budget,

fair, living lab, protection, and climate were with

higher tf-idf value and the terms budget and fair are

unique and not found in the other projects.

Data analysis suggests also that the gap between

DUT&SRIA and EN-UAC projects is within the

subject of energy, climate and ecosystem.

As the number of evaluated projects within the

EN-UAC programme rises, text analysis methods

will facilitate faster and more scalable comparative

assessments among them compared to traditional

approaches.

ACKNOWLEDGEMENTS

The research leading to these results has received

funding from the Horizon 2020 ERA-NET Cofund

Urban Accessibility and Connectivity (EN-UAC)

projects “Individual Mobility Budgets as a

Foundation for Social and Ethical Carbon Reduction”

(MyFairShare) and “Accessibility and Connectivity

knowledge hub for Urban Transformation in Europe”

(ACUTE).

FEMIB 2024 - 6th International Conference on Finance, Economics, Management and IT Business

REFERENCES

ACUTE (2023). JPI Urban Europe’s EN-UAC Knowledge

Hub Accessibility and Connectivity knowledge hub for

Urban Transformation in Europe. https://jpi-

urbaneurope.eu/news/the-knowledge-hub-a-platform-

for-gathering-knowledge-and-experiences-for-a-

sustainable-city/

Arhipova, I., Bumanis, N., Paura, L., Berzins, G., Erglis,

A., Rudloff, C., Vitols, G., Ansonska, E., Salajevs, V.

and Binde, J. (2023). Municipal Transport Route

Planning Based on Fair Mobility Budget. Rural

Sustainability Research, 50(345), 44-58.

https://doi.org/10.2478/plua-2023-0014

Arhipova, I., Bumanis, N., Paura, L., Berzins, G., Erglis,

A., Vitols, G., Ansonska, E., Salajevs, V., Binde, J.

(2023). Optimizing transport network to reduce

municipality mobility budget. Proceedings of the 5th

International Conference on Finance, Economics,

Management and IT Business (FEMIB 2023), 38–47.

Bartzokas-Tsiompras A. and Bakogiannis E. (2023).

Quantifying and visualizing the 15-Minute walkable

city concept across Europe: a multicriteria approach,

Journal of Maps, 19(1), https://doi.org/10.1080/174

45647.2022.2141143

Bartzokas-Tsiompras, A., Bakogiannis, E. and Nikitas A.

(2023). Global microscale walkability ratings and

rankings: A novel composite indicator for 59 European

city centres. Journal of Transport Geography, 111.

https://doi.org/10.1016/j.jtrangeo.2023.103645.

Basecamp (2023). JPI Urban Europe on Basecamp - a

space dedicated for exchanges with to with the first 15

EN-UAC Co-funded projects coordinators.

https://3.basecamp.com/3660243/projects

Clerici Maestosi, P., Andreucci, M.B. and Civiero, P.

(2021). Sustainable Urban Areas for 2030 in a Post-

COVID-19 Scenario: Focus on Innovative Research

and Funding Frameworks to Boost Transition towards

100 Positive Energy Districts and 100 Climate-Neutral

Cities. Energies, 14(1), 216. https://doi.org/10.3390/

en14010216

DUT (2023). JPI Urban Europe’s programme Driving

Urban Transitions to a Sustainable Future.

https://dutpartnership.eu

EN-UAC (2023). JPI Urban Europe’s ERA-NET Urban

Accessibility and Connectivity partnership. https://

www.era-learn.eu/network-information/networks/en-

uac

Feinerer, I, Hornik, K. (2020). tm: Text Mining Package. R

package version 0.7-8. https://CRAN.R-project.org/

package=tm

Gephi (2023). The Open Graph Viz Platform.

https://gephi.org/

Lancia, F. (2007). Word Co-Occurrence and Similarity in

Meaning. Some Methodological Issues. Linguistics,

Computer Science Corpus ID: 46954203,

https://api.semanticscholar.org/CorpusID:46954203

Moreno C. The 15 minutes-city: for a new chrono-

urbanism! (2019). http://www.moreno-web.net/the-15-

minutes-city-for-a-new-chrono-urbanism-pr-carlos-

moreno/.

MyFairShare (2023). Horizon 2020 ERA-NET Cofund

Urban Accessibility and Connectivity (EN-UAC)

project “Individual Mobility Budgets as a Foundation

for Social and Ethical Carbon Reduction”. https://jpi-

urbaneurope.eu/project/myfairshare/

Olsen, J. R., Nicholls, N., Caryl, F., Mendoza J. O., Panis,

L. I., Dons, E., Laeremans, M., Standaert, A., Lee, D.,

Avila-Palencia, I., de Nazelle, A., Nieuwenhuijsen M.

and Mitchell, R. (2022). Day-to-day intrapersonal

variability in mobility patterns and association with

perceived stress: A cross-sectional study using GPS

from 122 individuals in three European cities. SSM -

Population Health, 19. https://doi.org/10.1016/

j.ssmph.2022.101172.

R Core Team (2021). R: A language and environment for

statistical computing. R Foundation for Statistical

Computing, Vienna, Austria. https://www.R-

project.org/.

Robinson, D., Silge, J. (2017). Text Mining with R: A Tidy

Approach. United States: O'Reilly Media.

SRIA 2.0 (2023). JPI Urban Europe’s Strategic Research

and Innovation Agenda. https://jpi-urbaneurope.eu/

about/sria/sria-2-0/

Szmrecsanyi, B. (2012). Grammatical Variation in British

English Dialects: A Study in Corpus-Based

Dialectometry. Cambridge University Press.

Vijayarani, S., Ilamathi, M.J. and Nithya, M. (2015).

Preprocessing Techniques for Text Mining – An

Overview. International Journal of Computer Science

& Communication Networks, 5(1), 7–16.

Applying Text Analytics Methodology to Analyze Project Reports