|
|
Web application for keyword-aware walking route search
Zhukov, Dmitry ; Nečaský, Martin (advisor) ; Kopecký, Michal (referee)
Most mainstream web mapping applications implement location-based direction search. The typical workflow involves constructing an explicit sequence of places to visit. In this thesis, we aim to develop a web application that lets users formulate search queries in terms of categories, each composed of a keyword and attribute filters. A resulting route passes through at least one place from each category. The search procedure is formalized as a variant of the generalized Traveling Salesman Problem and solved with the help of polynomial-time heuristics. The application follows the three-tier architecture pattern. The frontend is imple- mented as a single-page application written in TypeScript using the React library, while the backend is programmed using the ASP.NET framework. We utilize the Open- StreetMap dataset and two knowledge graphs, Wikidata and DBPedia, as the basis for the conceptual model. Data is preprocessed and stored in MongoDB, which also serves as an efficient index. The OSRM routing engine helps calculate shortest paths and estimate network distances. Last but not least, the application stores user data in a decentralized way, either in IndexedDB or a Solid pod. The former is a standardized in-browser database, while the latter is part of an emerging technology that gives users control over...
|
|
|
Throughput optimization of a multistage data visualisation pipeline
Molnárová, Soňa ; Šmelko, Adam (advisor) ; Nečaský, Martin (referee)
Recent technological advances allowed vast increases of the size of datasets acquired by flow cytometry. With more cells and features captured, manual exploration and analysis of the data become challenging; multiple unsuper- vised methods were thus created to simplify the task. This thesis describes BlosSOM, a software that further improves the analysis possibilities by allow- ing interactive CUDA-accelerated supervision of the most common analysis approaches - dimensionality reduction, clustering, and annotation. The thesis overviews the features of BlosSOM implemented by the thesis author for the previous research project and details new developments, mainly a new dynamic workload balancing approach that allows BlosSOM to easily scale to very complex datasets while maintaining interactive framerate. To- gether with other improvements, BlosSOM is a production-ready software that offers all essential tools to analyze the flow cytometry data.
|
|
|
Application for management of map object collections
Bojačko, Martin ; Nečaský, Martin (advisor) ; Kopecký, Michal (referee)
The main goal of this thesis is to design, implement and test a web application that is used for searching and managing map objects. The application is inspired by successful projects such as Google Maps, OpenStreetMap and Mapy.cz, but it is also enriched with new functionalities. The application provides users with various options for searching map objects based on category selection, defining an area and application of different filters. The user can save map objects to one of his collections and subsequently man- age the saved map objects. For each saved map object, the application allows defining the visit time or adding notes. Additionally, the application provides information about individual map objects, such as an image, description, or a link to a Wikipedia article about the given object. The Wikidata project was chosen as the data source. Informa- tion about map objects is obtained from Wikidata using SPARQL queries. These queries were sent to the Wikidata Query Service through the SPARQLWrapper library, which also extracted the data. The acquired data was stored in a MySQL database. The appli- cation is implemented as a Single-page application, using the React library for the user interface and the Flask library for the server-side. The user interface of the application is written in TypeScript,...
|
|
|
Automatic schema extraction from RDF data
Kuželík, Jan ; Škoda, Petr (advisor) ; Nečaský, Martin (referee)
The Resource Description Framework (RDF) is a model for the representation of semantic data. RDF allows the storage of information without a fixed schema. This provides more flexibility but the lack of a fixed schema poses a significant entry barrier to the utilisation of the stored data. The SPARQL language is used for querying an RDF database. Several works exist in the domain of schema extraction from SPARQL end- points. Most tend to provide a visual representation of the schema, rather than an immediately usable output. Many of these solutions perform a very thorough and lengthy extraction unsuitable for a web application environment and some are not even available online. This thesis introduces TypeSPARQ, an open-source web application for ex- tracting schemata from SPARQL endpoints. TypeSPARQ creates a visualisation of the endpoint's schema and offers options for exporting it. TypeSPARQ in- tegrates with LDKit, which provides type-safe access to SPARQL endpoints for TypeScript applications. These tools combined offer TypeScript developers a seamless process from endpoint exploration to integrating the endpoint within their projects. 1
|
|
|
Web application for searching and editing basketball players in Wikidata
Fedák, Richard ; Nečaský, Martin (advisor) ; Svoboda, Martin (referee)
There are many ways to publish data in a machine-readable form on the Internet. One way is through entities and their interconnection, where two entities are interconnected if they have a relationship with each other. Such a data structure is referred to as a knowledge graph, which is used in the world of open data. The aim of this work is to create a web application that queries two such graphs, namely the knowledge graph Wikidata and DBpedia. The application is able to search for a basketball player and obtain data about him from the given knowledge graphs. It allows the user to discover various connections between the searched player and others using shared categories and simply add basic missing information using the application to the Wikidata knowledge graph.
|
|
|
Web application for searching recipes
Lhoťanová, Kristýna ; Nečaský, Martin (advisor) ; Peška, Ladislav (referee)
This thesis aims to develop a web application for searching recipes. The search for recipes is based on aggregating datasets from the existing recipe websites and extending the data using knowledge graphs. Knowledge graphs were represented by DBpedia and Wikidata projects. These were used to gather data about ingredients. Data were extracted using the Apify web scraping library and stored in the database system Apache CouchDB using the document model. The application provides the user with different options for filtering results, including faceted search. Faceted search is implemented using the Apache Solr platform. The focus is on searching based on ingredients. The web application is a single-page application implemented using the React.js library at the frontend and the Express.js framework at the backend. Both parts of the application are written in statically typed language TypeScript and exchange information through REST API. 1
|
|
|
Evolution Management in Multi-Model Databases
Bártík, Jáchym ; Holubová, Irena (advisor) ; Nečaský, Martin (referee)
Multi-model databases allow us to combine the advantages of various data models by storing different types of data in different models. However, this technology is still relatively immature, lacks standardization, and there are not any tools that would allow us to model multi-model data or manage their evolution. This thesis (i) provides an in- troduction to MM-cat, i.e., the framework for modeling multi-model data, (ii) describes the implementation of the framework, (iii) designs a workflow and a set of schema modi- fication operations to facilitate evolution management and (iv) performs experiments to prove their reliability and scalability. 1
|
|
|
Model-driven approach for data schema definitions modeling
Stenchlák, Štěpán ; Nečaský, Martin (advisor) ; Holubová, Irena (referee)
This work analyzes, formalizes, and implements a framework for multi-level concep- tual modeling of various serialization formats based on Model-Driven Architecture and previously developed tools XCase and eXolutio. It enables users to model their schema from a conceptual model in one general form from which multiple schema formats can be derived alongside documentation and transformation scripts. The thesis introduces base formalisms and findings and analyzes advanced requirements for the following work in this area, such as evolution and inheritance of schemas. The primary use case of the tool is modeling formal open standards for publishing open data for the government and public institutions of the Czech Republic. Nevertheless, the intent is to make the tool for general schema modeling. 1
|
|
|
Visualization of open data published according to the Formal Open Standard for public administration bulletin boards
Golubeva, Evgenia ; Klímek, Jakub (advisor) ; Nečaský, Martin (referee)
The Bachelor thesis is about the design and implementation of a web application which visualizes data from public administration bulletin boards. It uses open data published according to a new machine-readable format specified as a Formal Open Standard. The application is intended not only for users from the general public who can use it to search and filter information from bulletin boards but also for data publishers. The application performs validation of published data and clearly displays any deficiencies. To retrieve data the application uses the SPARQL endpoint of the National Data Catalog. It is implemented as a single-page application in TypeScript using React fra- mework. 1
|
|
|
Internet shop for government services
Abrahám, Milan ; Nečaský, Martin (advisor) ; Klímek, Jakub (referee)
Services provided by the state are currently not being presented to citizens in a uni- form way. The Register of Rights and Obligations (Registr práv a povinností) contains data about all provided services. The goal of this thesis is the creation of a functional internet eshop prototype, targeted to citizens. In addition to searching and recommen- ding, the eshop shall also allow the ordering of services and communication with autho- rities. This thesis describes the register and its purpose. The available data is analyzed. There are not enough of them for a proper functionality of some parts, so their adjust- ment is suggested. Communication with authorities isn't unified, this work presents a new way, using forms. They are generated dynamically, according to a designed data structure. The architecture and development of the eshop are described, including a testing process. The possibilities of future expansion are summarized. 1
|
guest ::
