Sie können Operatoren mit Ihrer Suchanfrage kombinieren, um diese noch präziser einzugrenzen. Klicken Sie auf den Suchoperator, um eine Erklärung seiner Funktionsweise anzuzeigen.
Findet Dokumente, in denen beide Begriffe in beliebiger Reihenfolge innerhalb von maximal n Worten zueinander stehen. Empfehlung: Wählen Sie zwischen 15 und 30 als maximale Wortanzahl (z.B. NEAR(hybrid, antrieb, 20)).
Findet Dokumente, in denen der Begriff in Wortvarianten vorkommt, wobei diese VOR, HINTER oder VOR und HINTER dem Suchbegriff anschließen können (z.B., leichtbau*, *leichtbau, *leichtbau*).
Dieses Kapitel befasst sich mit der Schaffung eines interaktiven Tools, das die strategische Planung im Eisenbahnsektor revolutionieren soll. Das im Rahmen des europäischen Projekts In2Smart2 entwickelte Tool beruht auf vier zentralen Säulen: Mess- und Überwachungssystemen, Datenmanagement, Tools zur Entscheidungsunterstützung und Technologiedemonstration. Das Hauptaugenmerk des Tools liegt auf der Optimierung der Ressourcen und der Verbesserung des Projektmanagements für Strukton Rail, einem Unternehmen, das an Erneuerungsprojekten und dem täglichen Projektmanagement beteiligt ist. Der Entwicklungsprozess gliedert sich in zwei Phasen. Die erste Phase umfasst die Auswahl von Power BI als Visualisierungstool, die Datenanalyse mittels Python und die Integration von GIS-Informationen. In der zweiten Phase werden dynamische Datenverbindungen, maschinelle Lernanwendungen, verbesserte GIS-Visualisierungen und Verbesserungen bei der Zugänglichkeit der Benutzer untersucht. Die interaktiven Funktionen des Tools, darunter Kapazitätskalender, Karten und Tabellen, bieten wertvolle Erkenntnisse für Entscheidungsfindung und organisatorische Aufgaben. Das Kapitel schließt mit der Hervorhebung des Potenzials des Werkzeugs für breitere Anwendungen über den Eisenbahnsektor hinaus.
KI-Generiert
Diese Zusammenfassung des Fachinhalts wurde mit Hilfe von KI generiert.
Abstract
An interactive tool for strategic planning is developed as part of the decision support and operational planning enhancements in the framework of the In2Smart2 project [1], working with Strukton Rail as the technology demonstrator.
The tool objective is to assist workers which form part of the decision process of strategic planning to visualize and analyze information, providing insights and useful conclusions in an interactive and straightforward way. The tool scope is the optimization of resources and company availability analysis to obtain forecasted workload optimization, the reduction of idle times and mobilization of adequate volumes of extra (external) capacity among others. It can be applied to a long-time horizon and to a short one.
The tool is implemented in Powe BI, elaborating methodologies and workflows for the implementation of the various interactive reports which support the decision-making and other managing and organizational tasks.
1 Introduction
In today's context, there exists a significant demand for a transformative shift in asset management, driven by innovative technologies, new economic possibilities, and enhanced legislative standards in the rail sector. This demand has propelled the In2Smart2 European project, aimed at intelligent asset management based on four key pillars: (1) Measuring and monitoring systems, (2) Data management, data mining and data analytics, (3) Decision support tools and systems and (4) Technology demonstration. These objectives led to the development of the interactive tool for strategic planning.
The outcomes pertaining to the creation of diverse methodologies, approaches, and conclusions analyzed and defined during the project are applicable to any other Use Case. However, the customized application resulting from the project is tailored to suit the needs of the company demonstrator, Strukton Rail, aligning with its organizational structure and databases.
Anzeige
Strukton Rail operates across renewal projects and daily project management, each with distinct tasks and structures. There is a pressing need for optimizing resources and conducting an analysis of company availability to achieve tender offer optimization, reduce idle times, and secure adequate volumes of external workforce, among other goals. This tool can be applied to both a long-term horizon, involving the analysis of years of data, and a short-term horizon, enhancing daily maintenance activities by considering factors such as resources, costs, penalties, and work conditions.
2 Tool Implementation
In this section the different developments are detailed. During the project consecution two stages can be differentiated, a first stage which englobed the development of the main functionalities of the tool, and a second stage in which due to the accessibility to new resources new research lines are investigated to enrich the developments.
2.1 First Stage
The first step is the selection of the visualization tool. Different possibilities were analyzed (Shiny, Dash, Django and Flask among others). Finally, Power BI, an interactive data visualization software product developed by Microsoft with primary focus on business intelligence, is selected as the most adequate choice. This software provides cloud-base business intelligence services along with data warehouse capabilities such as data preparation, discovery and the creation of interactive dashboards. One of the great advantages of the software is its compatibility with other platforms, allowing data management using as data source platforms as Azure, or the use of visualizations based on other open-source and private software such as Python, RStudio or ArcGIS.
The next step on the developments was the data analysis. Since the different analysis and studies were performed on the provided data, the tool usefulness is highly dependent on the volume and quality of data provided. During the first stage of the project the datasets contain information about the client’s track possessions, about Strukton Customer Relationship management offers and sales opportunities, the company available capacity per discipline and GIS information. To analyze the different information sources, they are loaded and filtered using Python in search of possible wrong data. The data is cleaned, calculations are made in reference to project duration, the information is categorized, and a daily capacity is calculated for the construction of capacity calendars.
Anzeige
The GIS information described in the previous section is provided as various file geodatabases. The open-source software QGIS is used to load the data, make selections of specific locations, perform intersections between map layers coming from different sources and exporting the data in the adequate format for its utilization in Python and Power BI.
For the GIS representation inside Power BI different possibilities are analyzed and due to the problems of ArcGIS Maps for Power BI to handle the high-volume target maps, Icon Map (from leaflet) is selected as the adequate visual object. The introduction of the GIS information into the Icon Map visual object has an intermediate step, in which the datasets are exported from the shape file as a.csv files using Well Known Text format through QGIS software. This process allows the GIS information to be imported to Power BI as a data table, an afterwards introduced in the Icon Map visual object creating interactive maps.
Once all the data is pre-processed, removing missing and unimportant datasets, the information from the different datasets is connected by their Sales ID. The information from the projects is used to calculate the amount of work or Capacity measured in required shifts with respect to time for Strukton and its competitors. Since the fixed capacities per discipline is also provided, the available capacity with respect to time is also calculated. The different links between datasets and further capacity calculations are also performed using Python.
Power BI is used to analyze, transform and connect the data to generate the visuals. The data previously pre-processed using python and QGIS is introduced in the software.
Depending on the type of target visualization, it may be necessary to unpivot columns to generate visual objects filtered by attribute, for example, to filter per enterprise and discipline. When different tables are used in the same visual object, it may be necessary to generate new columns identifying to what type of data they belong to.
A date table is generated to be linked with all possible datasets to ease the process and further connection in the different visual objects. In a similar way, tables with columns listing the existing track possessions and project IDs are created with the same goal. Columns have been added with color codes to improve the visualization experience through format options.
For the visualizations to perform adequately the different columns from the datasets must be linked through model relationships to connect the information from the different datasets that are needed. The tables created to filter (date, enterprises and disciplines) connect the different datasets. For the elaboration of the connection models, the cardinality and cross filter direction of the active connections are carefully selected.
The adequate columns must be added to the visual object fields, depending on each visual object selected and the desired results, this data must be introduced selecting the adequate aggregation method (or none). Also, the different filters are defined and implemented.
The output for this part of the project is the interactive tool implemented in Power BI. Strukton Rail has access to the tool through a URL. The analysis is focused on capacity, maps, calendars, tables and other visual objects that will help on the decision-making process and any organizational and planning tasks and is organized in the following report pages:
Sold capacity comparison: it provides at first glance an idea about the sold capacity of the company in different time periods in comparison with the volume sold by the competitors and the company resources, which aids the decision making relative to the acquisition of new projects and outsourcing.
Capacity calendars: this page provides capacity information for the decision-making relative to the project acquisition and outsourcing, but also, since it is daily focused, it supports the decision making relative to daily planning and internal organization.
Available capacity and absence: this approach helps to investigate the effect that the absence has on the available capacity, optimizing resources and improving the organization.
Map with budget and capacity information: this report page helps in internal organizational tasks and management and optimization of resources
Map with track possession information: this composition linked to the track possession information allow users to obtain useful information about capacity and localization that helps in internal organizational tasks and management and optimization of resources as well as to analyze the time dimension on the GIS study, providing position and capacity information with organizational purposes. Figure 1 shows this report page visualization.
Fig. 1.
Visualization of map with track possession information page. Generated with Microsoft Corporation. Microsoft Power BI [Software]. Retrieved fromhttps://powerbi.microsoft.com.
On the second stage, after the development and deployment of the tool, new possibilities presented due to a time extension on the project and changes in the data provision allowed the research of complementary lines.
On the first stage, the tool was fully implemented based on the time and resources which limited the subtask, and after its finalization additional activities were strategically planned on the second stage to enhance the aforementioned tool in a series of research lines: (1) Dynamic data, (2) Machine learning, (3) GIS approach and (4) Fast access and key insights.
All previous developments (first stage) were made on a static database, meaning that all data provided was immutable throughout the Project. On the second stage, one of the objectives was the connection to the new dynamic databases provided by Strukton Rail, to generate visualizations which adapt to real-life scenarios. The first step for the dynamic connection was the generation of a dataflow. A dataflow is a collection of tables that are created and managed in workspaces in the Power BI service [2]. The dataflow connects with a series of static a dynamic data sources through a unique gateway, generating a series of tables stores in the Power BI Service. This dataflow is used in Power BI desktop to load, transform, and model the data and generate the reports. The data is imported into Power BI desktop and a new model is generated with the dataflow information which is used to generate new report pages. The final.pbix file obtained can be used by any user to obtain up-to-date information visualized on the defined report pages. The possibility of uploading the report to a shared workspace was also analyzed, finding some limitation in terms of the Python and API connection functionalities.
On the machine learning research line, the possibilities of working with business intelligence and AI tools (Power BI and Python), not in a sequential but in an integrated way and what kind of developments would this connection allow in terms of machine learning analysis are studied. After the analysis Python is proved to be adequate for the generation of visual objects made from scratch using Python libraries instead of the built-in features from Power BI as well as for data transformation and advanced algorithms application.
Complex statistical models with supervised and unsupervised leaning methodologies are available within the available Python libraries within Power BI, as well as tools related to model selection, dimensionality reduction and time series analysis. Some of the algorithms that were highlighted during the study were the ones relative to learning methodologies: clustering methodologies, Discrimination analysis, Outlier detection, Decision Trees and Random Forest algorithms, Support Vector Machine and Neural Networks. As a prove of concept a Decision Tree is applied using sklearn library [3] to predict the state of a particular request (won or lost).
On the GIS approach research line, besides the creation of new report pages with maps displaying the information from the dataflow using leaflet, the ArcGIS for Power BI approach could be tested. The built-in visual object has the advantages of using ArcGIS elements, such as the use of basemaps, ArcGIS infographics, buffer/driver time, find similar tool, find places, selection tools and the addition of reference layers from ArcGIS. This last element is especially relevant since Strukton works with ArcGIS to manage their GIS information which means that introducing their ArcGIS account credentials they could include in Power BI any GIS information from their database.
Another key difference from the second stage is the access to a specific ArcGIS API tool which allows the download of the desired GIS information in different formats. A methodology to access the API is developed using parameters in Power BI database, to obtain the information from the API, interconnect it with the rest of the information and to allow its visualization in the report page. Since the credentials to ArcGIS API are restricted as an invited user, to access the information the URL has been used with an access token.
On the fast access and key insights side, the idea of the project is to generate tools that ease the access to information and their interpretation in a direct and user-friendly way on the day-to-day work. Two ways of improving this aspect have been defined: (1) Mobile phone layout for the original report which allow users to visualize all the previously defined report pages in a direct and comfortable way and (2) Generation of key insights specific reports, which only give key information oriented to a specific task or employee. As an example, one report is created using the same data model as the global model, but this report is oriented only for the daily management of high and medium engineers.
3 Conclusions
In this project an interactive tool for strategic planning is created to support the decision-making and other managing and organizational tasks for Strukton Rail. On the first stage of the project, the static datasets provided were adequately pre-processed, processed and linked to obtain all possible useful information through Python and QGIS. The processed information is later used in Power BI software to generate the final data models to create the tool, which is distributed in different report pages where a series of visual objects, results and filters provide the desired information in an interactive way.
On the second stage of the project different research lines are analyzed to work with dynamic data and Python for the creation of the data model in an integrated way. The possibilities of applying machine learning technics are studied, new functionalities are analyzed for the GIS visualizations and improvements on the ease of the tool use are implemented including mobile phone layouts and specific report pages for key insights.
This project results can be analyzed from two different perspectives: on one hand the creation of the customized tool for the end-user in a shared environment with all previously defined enhancements, and on the other hand the creation of the different methodologies, connections, approaches and conclusions that can be applied to any other Use Case.
Open Access This chapter is licensed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license and indicate if changes were made.
The images or other third party material in this chapter are included in the chapter's Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the chapter's Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder.
