Uporaba orodij umetne inteligence za analizo podatkov

Đukanović, Valentina

| | SLO | ENG | Cookies and privacy

Bigger font | Smaller font

First page > Show document

Show document

Title:	Uporaba orodij umetne inteligence za analizo podatkov
Authors:	ID Đukanović, Valentina (Author) ID Kljajić Borštnar, Mirjana (Mentor) More about this mentor...
Files:	UN_Dukanovic_Valentina_2025.pdf (2,89 MB) MD5: 69E1C749DE247C312729726BBEE312F1
Language:	Slovenian
Work type:	Bachelor thesis/paper
Typology:	2.11 - Undergraduate Thesis
Organization:	FOV - Faculty of Organizational Sciences in Kranj
Abstract:	V diplomskem delu smo raziskali, kako lahko različna analitična orodja – klasična (Microsoft Excel, Orange) in generativna (ChatGPT, Claude) – podprejo proces analize podatkov. Namen raziskave je bil preveriti, ali lahko generativna orodja, ki temeljijo na velikih jezikovnih modelih, zagotovijo enakovredno natančnost in uporabnost kot tradicionalna orodja, hkrati pa ponudijo prednosti v hitrosti in razumljivosti rezultatov. Metodološko smo izvedli eksperiment na javno dostopnih podatkih o povprečnih bruto plačah v javnem sektorju (2018). Vsi postopki so bili izvedeni z enakimi vhodnimi podatki in enakimi nalogami: izračun povprečja, standardnega odklona in koeficienta variacije, kategorizacija po frekvencah ter priprava vizualizacij (histogram, stolpčni graf). Rezultati so bili primerjani glede na natančnost, hitrost, jasnost interpretacij in uporabniško izkušnjo, ki smo jo ovrednotili tudi z Likertovo lestvico. Rezultati kažejo, da Excel in Orange zagotavljata deterministično natančnost in popolno ponovljivost, vendar sta počasnejša in zahtevnejša za manj izkušene uporabnike. Generativna orodja so bila najhitrejša (≈ 2 minuti do prvih rezultatov) ter najbolje ocenjena pri jasnosti interpretacij in enostavnosti uporabe. Kljub temu so se pri njih pojavila manjša odstopanja v standardnih odklonih in kategorizacijah. To potrjuje hipotezo, da so generativna orodja ob ustrezni pripravi vhodnih podatkov enakovredna klasičnim, pri čemer ponujajo dodatno prednost v hitrosti in dostopnosti. Zaključimo lahko, da optimalne rezultate zagotavlja hibridni pristop: klasična orodja so primerna za čiščenje, validacijo in sledljivost, generativna pa za hitro prototipiranje, narativne povzetke in podporo uporabnikom brez tehničnega predznanja. Praktična priporočila vključujejo standardizacijo podatkovnih tokov, validacijo ključnih rezultatov v Excelu ali Orangeu ter usposabljanje uporabnikov za kombinirano uporabo obeh pristopov.
Keywords:	umetna inteligenca, analiza podatkov, Excel, Orange, ChatGPT, Claude
Place of publishing:	Kranj
Year of publishing:	2025
PID:	20.500.12556/DKUM-95243
COBISS.SI-ID:	253878275
Publication date in DKUM:	20.10.2025
Views:	263
Downloads:	28
Metadata:
Categories:	FOV
:	Copy citation

Average score:	(0 votes)
Your score:	Voting is allowed only for logged in users.
Share:

Hover the mouse pointer over a document title to show the abstract or click on the title to get all document metadata.

Licences

License:	CC BY-NC-ND 4.0, Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International

Link:	http://creativecommons.org/licenses/by-nc-nd/4.0/
Description:	The most restrictive Creative Commons license. This only allows people to download and share the work for no commercial gain and for no other purposes.
Licensing start date:	09.09.2025

Secondary language

Language:	English
Title:	Using artificial intelligence tools for data analysis
Abstract:	In this bachelor’s thesis, we explored how different analytical tools – traditional (Microsoft Excel, Orange) and generative (ChatGPT, Claude) – can support the process of data analysis. The aim of the research was to examine whether generative tools based on large language models can provide accuracy and usability equivalent to traditional tools, while offering advantages in speed and clarity of results. Methodologically, an experiment was carried out using publicly available data on average gross salaries in the public sector (2018). All procedures were conducted with the same input data and identical tasks: calculating averages, standard deviations and coefficients of variation, classifying salary groups into frequency categories, and producing visualizations (histogram, bar chart). Results were compared based on accuracy, speed, clarity of interpretation, and user experience, which was also evaluated using a Likert scale. The findings show that Excel and Orange ensure deterministic accuracy and full reproducibility but are slower and more demanding for less experienced users. Generative tools proved to be the fastest (≈2 minutes to the first results) and received the highest scores for clarity of interpretation and ease of use. However, minor deviations appeared in standard deviations and classifications. This confirms the hypothesis that, with properly prepared input data, generative tools are equivalent to classical ones, while providing the additional advantages of speed and accessibility. We conclude that optimal results are achieved through a hybrid approach: traditional tools are suitable for data cleaning, validation, and traceability, while generative tools are best used for rapid prototyping, narrative summaries, and supporting users without technical expertise. Practical recommendations include the standardization of data workflows, validation of key results in Excel or Orange, and training users in the combined use of both approaches.
Keywords:	Artificial Intelligence, Data Analysis, Excel, Orange, ChatGP, Claude

Comments

Leave comment

You must log in to leave a comment.

Comments (0)

0 - 0 / 0

There are no comments!

Back