TSJ Dashboard

TSJ Dashboard enables interactive and real-time analysis of the users’ questions and counsellors’ answers in the online counselling service This is me, which is maintained by the National Institute of public health RS. Also, the data from Google Analytics can be analyzed in TSJ Dashboard. The tool is designed as a web application. This means that it can be used on any computer with a web browser and internet access. TSJ Dashboard offers different kinds of data analysis and visualizations. Partial analysis (aggregated data, tables) can be exported in .csv format (can be easily opened with MS Excel). Also, all microdata can be exported for further statistical analysis.

Sex education in Slovenian secondary schools

The scientific monography is based on the survey done in the Spring of 2012. The main research question relates to the needs of the integration of sex education in Slovenian high schools education programs, the knowledge of sexually transmitted diseases and contraception among students. The results are comparable with the previous studies. They indicate the need to include the contents related to sexuality in the educational process. The results also show a low level of knowledge of sexually transmitted diseases among students and highlight the importance of the public media in the imparting the contents related to sexuality.

Software for generating reports

The 100 pages long automatic generated report outlines the basic descriptive statistics related to the operation of the This is me online counselling service in a given period. The current structure of the document was designed in cooperation with the This is me editorial office. The hottest topics and statistics in which they want to have a better insight were selected. However, there are still possibilities of adding any supplements at any time. The computer program used to generate the automatic generated report is written in Sweave, which combines the programming language for statistical analysis, namely R, and Latex, the document mark-up language. The input is the entire forum database, and the output is the document in PDF format.

Evaluation of the web counselling service

The evaluation of the web counselling service This is me combines: (i) the analysis of data collected with Google Analytics; (ii) the analysis of the adolescents’ questions and counsellors’ answers; (iii) the analysis of the web survey data; and (iv) the analysis of qualitative interviews among the users of the web counselling service. Analyses have shown that most users who have posted at least one question in the web counselling service are females aged 15 to 17 years old. Most questions are related to relationships, sexuality and sexual health, physical health and mental health. The users recognized the service as a precious and easily accessible communication channel with the professionals from the fields of mental and physical health. The web counselling service was ranked among the most highly rated web sites in its category, according to the ACSI.

Identification of cancer-related messages in the online healthcare community

The project aimed to identify cancer-related messages in the online healthcare community MedOverNet among approx. 3 million messages posted between 2015 and 2019. Different machine learning algorithms (naive Bayes classification, support vector machine, decision trees and random forests) were used on lematized text with removed stop-words. The most efficient was the random forest model with 92% sensitivity and 89% specificity. Around 0.6% messages were about cancer – 2.9% within health consulting discussions, 1.9% within social support discussions and 0.5% within socializing discussions. The identified messages were considered in the sampling procedure to analyze the quality of cancer-related information in the online healthcare community MedOverNet.

Questions’ analysis in This is me for the 20-years-long period

All questions posted between 2001 and 2021 in the Web counselling service were analyzed. The characteristics of the questions (e.g., age, gender and content of questions) have been determined by hand since 2012. These were used to determine the content of the remaining questions posted before 2012 (supervised machine learning algorithms were used). Overall, the physical, mental, and sexual health questions are equally represented. However, the number of questions about physical health decreased in the later years while the number of questions about mental health remained unchanged. According to gender: questions about drugs, addiction and sexuality, are more common among boys, while questions about mental health, relationships and body are more commonly asked by girls.

2021-2024

Mentoring relationships in the processes of sci. collaboration /.../

Learn More
2020-2023

Comparison and eval. of different approaches to blockmodeling /.../

Learn More
2020-2021

Opportunities and dangers of online health communities for healthcare

Learn More
2017-2020

Blockmodeling multilevel and temporal networks

Learn More
2014-2016

Cooperation networks in Slovene science

Learn More
2014

Social influence in dynamic networks

Learn More

Videos

Can we improve social support for the elderly during COVID-19?
  • SciAni
  • Marjan Cugmas and Anuška Ferligoj

This research highlights which groups of elderly people have either non-existent or insufficient social support during the COVID-19 pandemic and are therefore more vulnerable. For these elderly people, sustainable care policy planning and crisis intervention planning should be organised especially for future waves of the coronavirus and other pandemics.

Statistics: doctoral degree programme
  • Laboratorij za multimedijo FE
  • Marjan Cugmas, Anuška Ferligoj and Bor Vratanar

The promotion spot highlights the PhD program in Statistics at the University of Ljubljana, aiming to attract both domestic and foreign students. Its primary objective is to arouse interest in the field of statistics and to present the wide range of specializations in the field of statistics. Two versions have been developed: a comprehensive 5-minute version and a concise 3-minute rendition.

Sometimes, you have to take a step back, to see the whole picture.

R Packages

nemBM

  • Author, Co-Creator

To study network evolution models and different blockmodeling approaches. Various functions enable generating (temporal) networks with a selected blockmodel type, taking into account selected local network mechanisms.

mri

  • Author, Creator

The package provides functions to compute the values of different modifications of the Rand and Wallace indices. The indices are used to measure the stability or similarity of two partitions obtained on two different sets of units with a non-empty intercept. Splitting and merging of clusters can (depends on the selected index) have a different effect on the value of the indices.

multiUS

  • Co-Author, Co-Creator

The package provides functions for multivariate analysis (e.g., factor analysis and discriminant analysis). It is primarily written for the course Multivariate analysis and for the course Computer intensive methods at the master's program of Applied Statistics at the University of Ljubljana.

dBlockmodeling

  • Contributor

The package contains functions to apply blockmodeling of signed (positive and negative weights are assigned to the links), one-mode and valued one-mode and two-mode (two sets of nodes are considered, e.g. employees and organizations) networks.

handyReport

  • Author, Creator

The package contains functions for summary tables and some statistical tests. Be aware of multitesting problem, when using this package.

ggthemeUL

  • Author, Creator

The package provides essential tools to produce charts that align with the comprehensive visual identity of the University of Ljubljana and its members. The rules defined by the comprehensive visual identity will be implemented starting January 1st, 2024. This package is designed to complement the functionality of ggplot2. See Tutorial.

Statistical Consulting

  • 2023
    Eva Gaberšek: Identification of autistic children with machine learning

    Mentor, Master's Thesis

  • 2023
    Vili Smolič: Linguistic analysis of suicide related questions in the online counselling service This is me

    Mentor, Master's Thesis

  • 2021
    Anja Ž. Kunaver: Multiple Testing Corrections: Case of Emissions Coupons' Price Movements

    Co-mentor, Master's Thesis

  • 2018
    Lovro Tacol: Interactions among physical development, motor skills and executive functions in children

    Consultant, Master's Thesis

  • 2017
    Martin Gec: How to improve competitiveness of disabled graduates of higher education in the labour market

    Consultant, Master's Thesis

  • 2016
    Anamarija Erčulj: The influence of learning aids on understanding art concepts in elementary school art education

    Consultant, Master's Thesis