• No results found

Ship and Weather Information Monitoring (SWIM) : Interactive Visulization of Weather and Ship Data

N/A
N/A
Protected

Academic year: 2021

Share "Ship and Weather Information Monitoring (SWIM) : Interactive Visulization of Weather and Ship Data"

Copied!
27
0
0

Loading.... (view fulltext now)

Full text

(1)LiU-ITN-TEK-A--09/012--SE. Visualization of Weather and Ship Data Oskar Eurenius Tobias Heldring 2009-03-18. Department of Science and Technology Linköping University SE-601 74 Norrköping, Sweden. Institutionen för teknik och naturvetenskap Linköpings Universitet 601 74 Norrköping.

(2) LiU-ITN-TEK-A--09/012--SE. Visualization of Weather and Ship Data Examensarbete utfört i vetenskaplig visualisering vid Tekniska Högskolan vid Linköpings universitet. Oskar Eurenius Tobias Heldring Handledare Patrik Lundblad Handledare Mikael Jern Examinator Mikael Jern Norrköping 2009-03-18.

(3) Upphovsrätt Detta dokument hålls tillgängligt på Internet – eller dess framtida ersättare – under en längre tid från publiceringsdatum under förutsättning att inga extraordinära omständigheter uppstår. Tillgång till dokumentet innebär tillstånd för var och en att läsa, ladda ner, skriva ut enstaka kopior för enskilt bruk och att använda det oförändrat för ickekommersiell forskning och för undervisning. Överföring av upphovsrätten vid en senare tidpunkt kan inte upphäva detta tillstånd. All annan användning av dokumentet kräver upphovsmannens medgivande. För att garantera äktheten, säkerheten och tillgängligheten finns det lösningar av teknisk och administrativ art. Upphovsmannens ideella rätt innefattar rätt att bli nämnd som upphovsman i den omfattning som god sed kräver vid användning av dokumentet på ovan beskrivna sätt samt skydd mot att dokumentet ändras eller presenteras i sådan form eller i sådant sammanhang som är kränkande för upphovsmannens litterära eller konstnärliga anseende eller egenart. För ytterligare information om Linköping University Electronic Press se förlagets hemsida http://www.ep.liu.se/ Copyright The publishers will keep this document online on the Internet - or its possible replacement - for a considerable time from the date of publication barring exceptional circumstances. The online availability of the document implies a permanent permission for anyone to read, to download, to print out single copies for your own use and to use it unchanged for any non-commercial research and educational purpose. Subsequent transfers of copyright cannot revoke this permission. All other uses of the document are conditional on the consent of the copyright owner. The publisher has taken technical and administrative measures to assure authenticity, security and accessibility. According to intellectual property law the author has the right to be mentioned when his/her work is accessed as described above and to be protected against infringement. For additional information about the Linköping University Electronic Press and its procedures for publication and for assurance of document integrity, please refer to its WWW home page: http://www.ep.liu.se/. © Oskar Eurenius, Tobias Heldring.

(4) Abstract This paper focus on the development of a tool for Ship and Weather Information Monitoring (SWIM) visualizing weather data combined with data from ship voyages. The project was done in close collaboration with the Swedish Meteorological and Hydrological Institute (SMHI) who also evaluated the result. The goal was to implement a tool which will help shipping companies to monitor their fleet and the weather development along planned routes and provide support for decisions regarding route choice and to evade hazard. A qualitative usability study was performed to gather insight about usability issues and to aid future development. Overall the result of the study was positive and the users felt that the tool would aid them in the daily work.. ii.

(5) Acknowledgements A special thanks to SMHI Global Product Manager Lennart Cederberg and NCVA [1] Professor Mikael Jern.. iii.

(6) Contents. 1 Introduction 1.1 Background . . . . . . . . . . . . . . . . . . . . . . . . . . 1.1.1 Swedish Meteorological and Hydrological Institute 1.1.2 Weather Introduction . . . . . . . . . . . . . . . . 1.2 Purpose . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1.3 Related Work . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . .. . . . . .. . . . . .. . . . . .. . . . . .. . . . . .. . . . . .. . . . . .. . . . . .. 1 1 1 2 3 4. 2 SWIM 2.1 Development Process . . . . . . . . . . 2.2 GeoAnalytics Visualization (GAV) . . 2.3 Data . . . . . . . . . . . . . . . . . . . 2.3.1 Weather . . . . . . . . . . . . . 2.3.2 Voyage . . . . . . . . . . . . . . 2.4 Visual Representation and Interaction 2.4.1 World Map . . . . . . . . . . . 2.4.2 Parallel Coordinates Plot . . . 2.4.3 Voyage Time Graph . . . . . . 2.4.4 Linked Views . . . . . . . . . . 2.5 Task Flow . . . . . . . . . . . . . . . .. . . . . . . . . . . .. . . . . . . . . . . .. . . . . . . . . . . .. . . . . . . . . . . .. . . . . . . . . . . .. . . . . . . . . . . .. . . . . . . . . . . .. . . . . . . . . . . .. . . . . . . . . . . .. 5 6 6 6 6 7 7 7 9 10 10 12. . . . .. 14 14 14 14 15. 3 Results 3.1 Usability Evaluation 3.1.1 Participants . 3.1.2 Procedure . . 3.1.3 Results . . .. . . . .. . . . .. . . . .. . . . .. . . . .. . . . .. . . . .. . . . .. . . . .. . . . .. . . . . . . . . . . .. . . . .. . . . . . . . . . . .. . . . .. . . . . . . . . . . .. . . . .. . . . . . . . . . . .. . . . .. . . . . . . . . . . .. . . . .. . . . . . . . . . . .. . . . .. . . . . . . . . . . .. . . . .. . . . . . . . . . . .. . . . .. . . . . . . . . . . .. . . . .. . . . . . . . . . . .. . . . .. . . . . . . . . . . .. . . . .. . . . .. . . . .. . . . .. . . . .. . . . .. . . . .. . . . .. . . . .. 4 Discussion 17 4.1 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17 4.2 Future Work . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18 4.2.1 Forecast Certainty . . . . . . . . . . . . . . . . . . . . . . . . . . . 18 Bibliography. 19. List of Figures. 21. i.

(7) 1 Introduction In this report chapter 1 will present background information to the thesis subject area together with an introduction to SMHI and the weather parameters used. This is followed by a description of the purpose of the thesis work along with a review of related work. Chapter 2 presents an overview of the usage of SWIM, describes the tools used during development, introduces the data sources and how the data is handled as well as a detailed description of the visual components in SWIM and how the interaction is linked between them. In chapter 3 the results are presented in form of a usability study. Lastly, chapter 4 consists of conclusions drawn and presents ideas for further work.. 1.1. Background. Visualizing multiple data sets from different sources represents a major challenge for the future. Using data fusion [2], [3] to visualize integrated spatial and temporal data it is possible to gain insight and retrieve information that would otherwise have to be obtained using much more time consuming methods. One subject area of particular interest is the shipping industry where the need of being able to draw accurate conclusions and make well planned decisions regarding routes and weather factors are important. There are a large number of variables affecting the success of every sea voyage performed by vessels such as cargo ships. Important aspects being considered are speed, safety and costs because of their direct impact on the voyage performance. Weather factors are the main concern when considering these aspects and therefore shipping companies consult meteorologists about weather information and route optimization.. 1.1.1. Swedish Meteorological and Hydrological Institute. SMHI’s mission is to manage and develop information on weather, water and climate that provides knowledge and advanced decision-making data for public services, the private sector and the general public. SMHI is a government agency under the Ministry of the Environmental and offers products and services that provides various kinds of enterprises and organizations with an important foundation for decision-making where general forecasts and weather warnings, industry-specific services, simulations and analyzes, statistics, climate studies and contracted research are some examples [4]. During our work we have been positioned at the Business and Media Services Department which markets and produces customized and industry-specific forecasting and data services aimed at customers in spheres such as media, industry and commerce, agriculture, the private market, shipping, land-based transport, energy and the building trade 1.

(8) 1.1. BACKGROUND. 2. and property management. The Shipping Department at the Business and Media Services Department uses a weather routing system to monitor voyages according to weather criteria, voyage data and ship performance to support captains decision-making regarding route optimization and harsh weather. This weather routing system enables shipping companies with their respective captains to receive updated weather forecast information and guidance. In situations when shipping companies prefer to monitor their voyages themselves a web-based system called FleetWeb is used which provides similar, but more limited, functionality as the weather routing system.. 1.1.2. Weather Introduction. Weather parameters are not always easy interpreted but important aspects to understand and therefore follows a short description of the most common parameters used in SWIM. Waves The wave height parameter is separated into three different types: significant wave height, wind wave height and swell height. Significant wave height is the average wave height, trough to crest, of the one-third largest waves and was intended to mathematically express the height estimated by a trained observer. It is today commonly used as a measure of the height of ocean waves. Wind wave height is the wave height produced by the local wind and swell height is the wave height generated by earlier formations of wind systems in different geographical locations. The wave period parameter and wave direction parameter are also categorized according to significant waves, wind waves and swell where the period is the unit time between two crests and the direction the path the wave travels. Wave spectral kurtosis or freak wave index is a wave parameter describing the probability of the generation of waves with extreme heights compared to the mean wave height in that area. In probability theory and statistics, kurtosis is a measure of the ”peakedness” of the probability distribution of a real-valued random variable, in this case the significant wave height. Wind The wind parameter is commonly classified by its spatial scale, wind speed and the geographic regions in which it occur. Wind direction is reported by the direction from which it originates, e.g. a northerly wind blows from the north to the south. Wind speed is the movement of air in an atmosphere and is a scalar quantity, the magnitude of the vector of motion (figure 1.1). Wind measured at level 89 of an isobaric scale, approximately 40 meters above sea level, is a suitable wind parameter when evaluating shipping conditions.. Figure 1.1: Wind speed and direction is usually represented in visualizations by wind flags in unit knots..

(9) 1.2. PURPOSE. 3. Air Pressure The force per unit area exerted against a surface by the weight of air above that surface at any given point in the earth’s atmosphere is a common definition of atmospheric pressure. Pressure decreases with increasing elevation where low pressure areas have less atmospheric mass above their location and pressure increases with decreasing elevation when high pressure areas have more atmospheric mass above their location. Air pressure measured at sea level is referred to as the mean sea level pressure (MSLP) and is the pressure normally given in weather reports. Using MSLP, pressure considered high pressure or low pressure do not depend on geographical location and the reduction to sea level normalizes fluctuations in pressure which makes isobars on weather maps meaningful and useful tools. Temperature The temperature parameter suitable when evaluating shipping conditions is the 2 meter temperature, measured two meters above sea level, which is given in Celsius. Cloud Cover Cloud cover refers to the fraction of the sky obscured by clouds when observed from a particular location and its parameter is given in unit percent. Precipitation Precipitation is any product of the condensation of atmospheric water vapor that is deposited on the earth’s surface. The precipitation parameter describes either rain or snow in their corresponding height of water measured in meters.. 1.2. Purpose. FleetWeb suffers from design flaws such as low interactivity, cluttered displays, unstable initialization and lack of filtering possibilities. SWIM was created as a prototype for a future version of FleetWeb to aid shipping companies monitoring vessels according to forecasted weather development along planned routes (figure 2.1). The tool provides support for decisions regarding route choice with respect to future weather development and warnings of extreme weather factors. With weather forecasts and voyage information provided by the European Centre for Medium-Range Weather Forecasts (ECMWF) [5] and SMHI [4] shipping companies can monitor fleets using highly interactive visual representations visualizing weather parameters and detailed voyage information. Using a geographic map visualization together with a parallel coordinates plot (PCP) brings great possibilities when exploring the data. SWIM was designed with the following tasks in mind: • Detailed monitoring of voyages based on planned and reported waypoints. • Easy exploration of weather forecast according to both geographic positions and along routes..

(10) 1.3. RELATED WORK. 4. • Using a PCP together with a geographic map visualization to find interesting voyages according to weather parameters. • Using highly interactive linked visual representations to facilitate exploration of data.. 1.3. Related Work. Visualization of spatio-temporal data has been the subject of several recent research papers, see [6] for an overview. GeoVista Studio [7] and CommonGIS [8] are general systems which support exploratory data analysis with decision-making and Andrienko and Andrienko has illustrated other motivating approaches in earlier papers [9], [10]. According to Jern and Franzn most of these systems lack support for analyzing simultaneously multiple attributes data and spatio-temporal behavior [11] and as a response a generic ”GeoAnalytics” visualization (GAV) [12] toolkit is introduced. Jern et al. emphasize the advantages and need of multiple linked views (MLVs) for dynamically exploring timevarying, geographically referenced and multivariate attribute data. Parallel coordinates [13], [14], [15], [16] with embedded visual inquiry methods that serve as a visual control panel for dynamically linked and coordinated views is a well-known visualization technique for multivariate data [17]. In VISPER [18] multi-dimensional, multi source, time-varying and geospatial information from voyage analysis is represented to facilitate decision-making. Therefore Lundblad et al. propose the use of common InfoViz visual representations for multivariate data together with geographic mapping to aid users in their work examining voyage performance. Roberts [19] provides a review of multiple linked-view tools, methodologies and models, discusses related challenges and ideas, and provides some rudiments for coordination within a geovisualization context. In accordance with Roberts linking and relating information in one view to that of other views will assist the user in the exploration process and may provide additional insight into the underlying information. Operations such as filtering, dynamic queries [20] and selections applied simultaneously in different views when affecting not only the same information but more effectively collections of different data is pointed out by Roberts as more interesting and an important aspect to consider. Spirkovska and Lodha presented AWE - Aviation Weather Data Visualization Environment [21] which was designed to meet an urgent need of critical weather element visualizations. AWE focuses on interactive graphical displays of these weather elements such as, meteorological observations and terminal area forecasts and maps them onto a cartographic grid specific to the pilot’s interest. Song et al have designed an integrated atmospheric visual analysis and exploration system for interactive analysis of weather data sets [22]. These tools provide meteorologists with new abilities to analyze their data and answer questions on regions of interests, ranging from physics-based atmospheric rendering to illustrative rendering containing particles and glyphs. In accordance with Dorneich et al. challenges lie in effectively integrating voyage and weather information in the same tool to facilitate decision-making [23]..

(11) 2 SWIM SWIM is developed using the GeoAnalytics Visualization (GAV) Framework [24] and is a tool shipping companies can use to monitor voyages and receive up to date weather forecasts corresponding to their routes. SWIM will on startup automatically find the latest weather forecast delivered by the ECMWF. In SWIM the user is presented with three different modes accessible using tabs. In the first mode a selection of static weather parameters are presented in a weather parameter menu sorted according to priority level, e.g. basic weather, advanced weather and extra weather (figure 2.1). The second mode presents a searchable list menu with all active voyages retrieved from the SMHI database. In the third mode focus lies on one or more selected voyages with detailed information presented.. Figure 2.1: The SWIM overview. In the world map ships are visualized using glyphs and their routes are plotted as lines. Significant wave height is displayed using an iso-surface where warmer color indicates higher waves. The current time step can be changed using the time slider positioned underneath the world map. Using the weather parameter menu to the left the user can select which parameter to visualize and which representation to use. At the bottom weather parameters are visualized in the parallel coordinates plot.. 5.

(12) 2.1. DEVELOPMENT PROCESS. 2.1. 6. Development Process. The first part of the thesis work was devoted to creating a reader for weather data. Developing a flexible reader for the binary weather data (described in section 2.3.1) was time consuming yet rewarding later on. With the ability to read and store weather parameter data focus shifted to develop map visualizations such as iso-surfaces, iso-lines, wind flags, arrow glyphs, number glyphs and freak wave glyphs. During the next step work was put into expanding the underlying data flow to store data over time and use this to enable the visualization of time dimension in the world map view. Next a reader for voyage data (section 2.3.2) was created followed by map visualizations of ships and routes. With the voyage data available weather data could be linked to voyages and the voyage time graph was implemented followed by the PCP. For the last part of the thesis work focus where on data flow and linking between views, interaction and interface design.. 2.2. GeoAnalytics Visualization (GAV). GAV [24] is a component toolkit for dynamically exploring time-varying, geographically referenced and multivariate attributes simultaneously and to enable the capture of the interactive visual process into information packages that allow the analysts to communicate their discovery and decision recommendations. GAV includes components based on a synergy of technologies from information visualization (InfoViz), geovisualization (GeoViz) and scientific visualization (SciViz). The GAV components are developed in C# based on Microsoft’s low-level DirectX graphics library and fulfill many generic requirements for a Visual Analytics application design framework.. 2.3 2.3.1. Data Weather. The input weather data used by SWIM is in GRIB (Gridded Binary) [25] format which is a mathematically concise data format commonly used in meteorology to store weather data. It is developed and standardized by the World Meteorological Organization WMO. The GRIB format is used world-wide by meteorological centers for numerical weather prediction output and serves as an efficient solution of transferring large volumes of digital gridded weather data. In our work focus lies on GRIB files originated from the European Centre for MediumRange Weather Forecasts (ECMWF) [5]. The first part of our thesis work consisted of creating a reader for GRIB files. By methodically reading segment by segment using the GRIB API [26] as reference, the GRIB reader developed from a simple read and print functionality to a fully developed binary reader being able to collect and store data for specific parameters as well as supporting different types of GRIB files. In a GRIB file the parameters are positioned sequentially in a binary data stream. Each parameter segment consists of the standard GRIB file fields; including a parameter description, measurement level, grid description and the actual data. The GRIB files used by SWIM are of data grid resolution 1,5 degrees in both latitude and longitude.

(13) 2.4. VISUAL REPRESENTATION AND INTERACTION. 7. which results in 240 horizontal and 120 vertical points summed up to a total of 29040 data points in the world map. SWIM allows for exploration of weather forecasts over a ten days period. Each GRIB file consists of a selection of weather parameters with corresponding weather data for a certain date and time. SWIM are using a total of 31 GRIB files with a forecast time interval of 6 hours the first 5 days and an interval of 12 hours the later 5 days. When a weather parameter is selected in the weather parameter menu the parameter is located and read from each of the 31 GRIB files and saved into a data array which enables the time dimension to be explored. Because of the large amount of binary data being interpreted during reading the process has been assigned its own thread in the system which enables simultaneously interaction with SWIM.. 2.3.2. Voyage. Voyage data is collected from a database provided by SMHI where extensive information regarding ships and their corresponding routes are available. During start up SWIM automatically reads all currently active voyages, which allow ships and routes to be visualized in the world map. Data retrieved for each voyage includes ship information such as operator, ship name and ship type. Voyage specific information includes estimated time of departure and estimated time of arrival, port of departure and port of arrival as well as a list of waypoint positions presenting the planned route of the voyage. Lastly, reports delivered by active are retrieved including current positions, speed and performance as well as weather observations at reported positions. Reported positions are combined with planned waypoints to form a list, with latitude and longitude positions describing voyage routes, used to visualize voyages routes in the world map view. The position of ships at different time steps are estimated via interpolation where the time of last reported position, the time at selected time step and the estimated time of arrival are used to estimate the distance traveled between the last reported position and the position of the arrival port.. 2.4. Visual Representation and Interaction. In SWIM the users are given the possibility to study weather data cohering with voyage information in different views. A view represents a visual representation displaying selectable data with possibilities of interaction and filtering.. 2.4.1. World Map. The main view is the world map where all weather parameters originating from corresponding GRIB files can be presented according to selection in the weather menu. Options are provided on how to visualize a parameter, e.g. iso-line, iso-surface, glyph or numerical values. Routes with their corresponding ships are also visualized in the world map. The ships are represented by ship glyphs which are positioned and rotated according to an interpolation based on their pre-planned waypoints, reported positions and present time (figure 2.2). For these features to be useful it is necessary to look at not only the spatial variables but also temporal aspects. SWIM reads and store a ten days forecast consisting of 31 binary GRIB files which enables the temporal aspect to be.

(14) 2.4. VISUAL REPRESENTATION AND INTERACTION. 8. Figure 2.2: World map view visualizing significant wave height and wind. The significant wave height is visualized using an iso-surface where warmer color indicates higher waves. The wind is represented by wind arrow glyphs where the orientation shows the wind direction and the glyph shows the wind strength. Ships are shown using a ship glyph and their routes are drawn with lines.. taken into account and is represented by a forecast time slider. This slider is positioned at the first forecast prediction and with a sliding interval of six hours the first five days and a sliding interval of twelve hours the later five days it effectively enables the user to view weather changes and ship movement over time (figure 2.3). Routes with corresponding ships that at a specific time index has reached its final destination is marked as inactive and grayed out. The world map view with selected weather data visualized and chosen voyages plotted will work as an effective support to monitor several vessels simultaneously over a period of time. To facilitate usage of the world map view and maximize performance a series of graphics optimizations has been implemented. With a resolution of 240 longitude and 120 latitude grid points it is necessary to use a level of detail functionality to keep the world map useful and not cluttered when changing the map zoom. With a maximum of 29 040 data points on the data grid we use eight levels of detail based on a mean value calculation with respect to all interesting neighbors of each visible grid point (figure 2.4). Another important optimization feature is the active rectangular window we use to keep track of the current visible part of the world map. Using this functionality gives a performance advantage where weather parameter visualizations are only calculated and rendered if inside the active rectangular window. When zoomed in this optimization saves computational power and speeds up the interaction..

(15) 2.4. VISUAL REPRESENTATION AND INTERACTION. 9. Figure 2.3: The left side shows wind wave height as an iso-surface, ships visualized using ship glyphs and their routes as lines. The right side shows the same map area with the time slider advanced 48 hours.. Figure 2.4: The left side shows significant wave height as an iso-surface and mean wave direction as a regularly spaced arrow glyph. The right side shows the greater detail of the glyph representation received when zoomed in.. 2.4.2. Parallel Coordinates Plot. General Description PCPs consists of N equidistant vertical axes which corresponds to the selected variables (attributes). The axes are scaled to the [minimum, maximum] - range of the corresponding attribute and every data item corresponds to a line which intersects each of the axes at the point which corresponds to the value for the attribute. The main purpose of PCPs is to view relations in multivariate data and suitable data types supported are quantitative data and ordinal data. Highly interactive visualization applications can be developed using PCPs with techniques such as; brushing, highlighting (picking), dynamic filtering and color coding. Implementation The PCP view displays all the voyages with loaded parameters for the selected time step. Each voyage is represented as a line in the PCP where each axis corresponds to a specific.

(16) 2.4. VISUAL REPRESENTATION AND INTERACTION. 10. weather parameter (figure 2.5). This view gives the user an overview of the weather for all the voyages at a specific time step. Advancing the time slider will update the PCP with new data, namely weather data for each voyage at the new time step. Lines are colored according to the parameter value of the selected axis. Which parameter to use for coloring can be chosen by clicking the header the parameter axes. The colors of the lines are linked to the color of the ships in the world map view.. Figure 2.5: Time step parallel coordinates plot. The lines can be selected, filtered by moving the sliders on each axis and colored by clicking the header of the parameter.. 2.4.3. Voyage Time Graph. General Description Time graphs consists of N equidistant vertical axes which corresponds to sequential time steps in a given interval. Each attribute item is represented by a line which intersects each of the time axes at a point corresponding to the value at that time step. Using time graphs it is easy to visualize multiple attributes over time and observe temporal changes in data. Implementation When in the voyage information mode the voyage time graph gives the user an overview of how the weather will develop along the planned route of the selected voyage without the need of moving the time slider. Each parameter is represented as a line in the voyage time graph and each vertical axis represents a different time index according to additional hours added to the forecast creation date (figure 6). Each parameter is normalized according to predefined values in order to enable the different parameters to be plotted together. Straight lines near the bottom of the plot indicate stable and calm weather along the route for the forecast period. Curvy lines however indicate varying weather conditions and high peaks indicate extreme weather parameter values.. 2.4.4. Linked Views. At start up all voyages are displayed in the world map which provides an overview of routes and ship positions. The ability to highlight one or several voyages is important when exploring data and is implemented. In the world map view all the voyages except those highlighted will be slightly transparent preventing interference with highlighted voyages (figure 2.7). In the PCP lines of highlighted voyages will be selected. In the.

(17) 2.4. VISUAL REPRESENTATION AND INTERACTION. 11. Figure 2.6: Ship time graph visualizing weather parameters along the selected route. Each line represents a weather parameter where the x axes represents time and the y axes represents parameter value. The static coloring of the lines are predefined for each weather parameter.. voyage information tab text information will be presented for highlighted voyages and the voyage time graph will display data for one of the highlighted voyages.. Figure 2.7: Highlighting in the World Map view. The left side shows the map before selection and the right side after.. Different approaches for how to highlight voyages are presented to meet different user needs. An interesting voyage can for example be selected by clicking on it in the world map. To highlight a voyage according to name the ship list with its search function can be used. The PCP which presents weather for all voyages for a specific time step can be used to spot voyages experiencing harsh weather conditions and by clicking on a line the corresponding voyage will be highlighted. When a specific geographic area is of interest the selection area tool can be used. When the selection area tool is activated the user is presented with a circular area on the world map positioned according to mouse position. The size of the selection area can be manually set and when clicking a position in the world map all voyages traveling through that area at any point during the forecast period will be highlighted (figure 2.8). Hiding uninteresting voyages is achieved by using the filter sliders in the PCP. Filtered lines will be completely hidden as well as the corresponding voyages in the world map view. For example, the user can adjust the filters to only show voyages whose ships are exposed to a significant wave height above 5 meters (figure 2.9). When advancing the time slider the filter value is saved and only voyages whose ships are experiencing a significant wave height above 5 meters will be displayed at each time step..

(18) 2.5. TASK FLOW. 12. Figure 2.8: Selection area tool. The left side show the world map view with ships and routes where the grey circle is the selection area tool. When using the selection area tool ships that travel through the selected area at any time during the forecast period will be highlighted as shown on the right side.. Figure 2.9: Filtering using time step parallel coordinates plot. In the world map significant wave height is visualized using an iso-surface. The left side shows parallel coordinates plot with corresponding map before filtering. The right side shows filtering of ships according to significant wave height.. 2.5. Task Flow. A typical usage scenario in SWIM could be that the user is interested in ships experiencing harsh weather in the Atlantic outside of the English Channel. At startup the user will be presented with all voyages from the first time step displayed on the world map and the PCP will be empty since no weather data has been read. The first user action would be to read weather data of interest which for example could be significant wave height and wind. If the weather data is selected using the weather.

(19) 2.5. TASK FLOW. 13. menu data will be visualized on the map and displayed in the PCP. Focusing on ships experiencing harsh weather would be done by filtering out ships according to weather criteria using the PCP where ships not fulfilling the selected weather criteria disappears from the map. To further narrow the search according to geographic area the selection area tool would be used to highlight the area outside the English Channel. Ships on the world map not travelling through the selected geographical area during the forecast period will be grayed out. Detailed information about the filtered ships experiencing harsh weather in the Atlantic outside of the English Channel can now be retrieved by clicking on each preferred ship.. Figure 2.10: Which ships will experience harsh weather in the Atlantic outside the English Channel? A typical usage scenario in SWIM described using a task flow..

(20) 3 Results Since the product of our thesis work is a software tool we chose to evaluate the result using a usability study. Allowing users to give feedback on usability, design and techniques, conclusions could be drawn about both individual parts and the complete work.. 3.1. Usability Evaluation. The purpose of the evaluation was to explore qualitative usability issues with respect to visual representation and interactive representation [27] and to gather opinions regarding potential use and future development without intentions to do measurements.. 3.1.1. Participants. A selection of marine meteorologists and master students from the Media Technology and Engineering program [28] formed a group of six participants. The marine meteorologists had knowledge and experience in meteorology and ship routing and the MSc students had knowledge in information visualization, usability and design. Their occupations and respective domain knowledge was assumed to give insightful information regarding usability issues of SWIM.. 3.1.2. Procedure. Each session lasted nearly an hour. First the participant was given a demonstration of the tool and its functionalities. This was done to minimize the risk of misinterpretation as well as first time user misses since the intended users are a small number of people using the tool regularly. For the same reason the participants were given answers and explanations to all their questions about SWIM and its functionality during the test. After the demonstration the participant was given the control of the tool and was encouraged to explore some of the functionality earlier demonstrated. A number of tasks that was prepared beforehand were presented to the user in random order. The intention was not to evaluate the participant’s success of the specific tasks but simply to guide the user and encourage exploration. Throughout the evaluation a guide sat beside the participant actively discussing thoughts and opinions regarding the tool and a third person documented the important parts of the discussion. The last part of the evaluation consisted of a general discussion based around the criteria stated by Freitas et al. [27]. Visual Representations were discussed according to the following selection of criteria: Data density, limitations, visual/spatial orientation, data dimensions, relevant information, reference context, occlusion, display of detail and logical order. 14.

(21) 3.1. USABILITY EVALUATION. 15. Interactive Representations were discussed according to the following selection of criteria: filtering, selection of objects, search and querying, control of level of detail and viewpoint manipulation.. 3.1.3. Results. The marine meteorologists had limited previous experience working with visualization tools in comparison with the MSc students which were skilled using these tools. On the other hand the marine meteorologists had significant experience with weather data and route optimization tools and these differences were observed when studying the results of the usability evaluation. All participants completed the prepared tasks without major complications and the overall opinions describe SWIM as a tool with many possibilities. Direct data manipulation and filtering based on dynamic queries immediately changing the displayed view and additional linked views was greatly appreciated. All participants and especially the marine meteorologists described the possibility to filter voyages in the world map view according to criteria in the PCP as useful. During the evaluation the participants also pointed out both known and previously unknown issues and discussions about certain criteria regarding visual and interactive representation brought to light strengths and weaknesses of the tool. With regard to the visual representations, participants experienced the built in levelof-detail functionality in the map view as smooth and natural. However, comments where made about not being able to connect weather parameters to a specific route instead of regularly spaced grid-points. All found the spatial organization of visual representations as well thought-out and easy to interpret. The ability to choose different visualization methods for certain weather parameters in the world map and therefore being able to combine different representations was pointed out as well designed and important. The marine meteorologists requested functionality where the user would be automatically warned about voyages whose ships are affected by extreme weather situations rather than having to gain that insight during manual exploration. All participants agreed that the information displayed in the map view was relevant but additional information was requested in specific situations such as more detailed voyage information in the ship tool tip. With regard to the interactive representation a difference was noticed between the marine meteorologists and the MSc students in discovering interaction possibilities. The MSc students which are skilled using visualization techniques based on brushing, picking and filtering found it easier to discover and use these features. The marine meteorologists requested additional descriptions of the visual representations. However, even though the chosen target group is a small number of people using the tool regularly we found it possible for people without previous experience of similar techniques to embrace the fundamental ideas with a short introduction. All participants pointed out the difficulty of comparing weather values for different time steps. One participant suggested maximum and minimum axis values of the Time Step Coordinates Plot spanning over the entire forecast period as opposed to being time step specific. Participants also described the variety of selection methods to highlight a unique voyage as well implemented and the different approaches as good complements to each other. The Selection Area tool was especially appreciated by the marine meteo-.

(22) 3.1. USABILITY EVALUATION. 16. rologists who saw advantages with the tool enabling exploration of geographical areas. The voyage time graph was described by all as a powerful technique to quickly gain insight of weather development along a specific route. Comments were however expressed regarding carefully choosing normalization values for each parameter to avoid misinterpretation. Participants also requested an explanation of the colors of weather lines in the voyage time graph. The time slider was described by all participants as natural and appeasing to use. A more distinct representation of elapsed hours of the weather forecast was pointed out as a valuable addition. When selecting a single voyage requests about a go-to-ship functionality was expressed, automatically focusing and positioning the view over the ship corresponding to the selected voyage in the world map. Suggestions of improvements and additions to the GUI were given. One of those, a feature pointed out missing by the majority of the participants, was the ability to undo steps in the interaction..

(23) 4 Discussion 4.1. Conclusions. This report presents SWIM which is a tool to explore multi-source, time-varying and geospatial data. SWIM allows weather forecast data and voyage information to be analyzed using an interactive map view linked to a PCP and a time graph. Using SWIM analysts at shipping companies can monitor ship voyages and weather development along planned routes. Voyages whose ships are being exposed to harsh weather can easily be brought to focus and detailed route specific weather and voyage information can be retrieved. A qualitative usability evaluation was performed with a selection of six participants. Discussing and exploring the tool with each participant resulted in useful feedback regarding functionality and design issues. All participants completed the prepared tasks without major complications and the overall opinions describes SWIM as a tool with many advantages and interesting functionalities compared to similar tools in the same domain. In summary, especially appreciated features were: • The world map view with interaction possibilities such as highlighting, smooth zooming, brushing and weather parameter visualization options. • The PCP with the possibility to filter and distinguish voyages according to weather criteria, presenting results immediately in the world map view. • The voyage time graph which gives an overview of weather development along planned routes of selected voyages. • The ability to view and interact with the geospatial data in linked views with several highlighting possibilities. Using a PCP proved to be a powerful tool to filter and explore data in the world map view. Discovering voyages which fulfills certain weather criteria is an important feature which is made possible using the filtering and picking techniques of the PCP. The synergy between geovisualization (map) and information visualization (PCP) was demonstrated and evaluated as useful. The both known and previously unknown issues regarding visual and interactive representation that brought to light strengths and weaknesses during the evaluation will constitute the foundation for further development of SWIM.. 17.

(24) 4.2. FUTURE WORK. 4.2. 18. Future Work. For future development of SWIM more automation could be built into the tool. The tool could present a list with voyages meeting certain criteria for harsh weather conditions. Using this feature investigation of risky voyages could start immediately without the need of first filtering according to criteria. To easily compare weather values of different time steps for all voyages a second time graph could be implemented where weather development over time could be shown displaying one weather parameter and all or a selection of voyages. This would facilitate discovering of voyages experiencing the most extreme weather parameter value during the entire forecast period. The ability to use presets of user customizations, filtering criteria and standard parameters to read would increase the dynamics and speed when using the tool. Shipping companies regularly evaluates the performance of their ships by evaluating reports along the routes, weather factors and other variables of interest to determine if the voyage was performed as planned [18]. To facilitate this work SWIM could be extended to support post voyage analysis. Using the time slider to view past time steps and presenting reported data and way point specific weather SWIM could be a useful tool for this purpose.. 4.2.1. Forecast Certainty. Weather forecasts always come with a bit of uncertainty. The further into the future, the higher level of uncertainty. This is natural since exactly predicting and simulating the earth’s atmosphere is an extremely difficult task. Applications similar to SWIM have used only data from the latest available forecast which of course provide the highest levels of accuracy but still produces errors, especially with long predictions. A time-lagged forecast [29] is the result of weighing earlier forecasts according to specified values and using these lagged forecasts the user would be able to estimate the accuracy of the current forecast. This lagged forecast would behave with inertia because of its dependence on earlier forecasts and would often be a bit too conservative when evaluating the forecast. However, comparing the lagged forecast with the latest forecast would give the user a sense of the certainty of the forecast. If the lagged forecast present similar values as the latest single forecast it indicates that the prediction is reliable. On the other hand, if the two values differ greatly it is an indication of uncertainty in the forecast. This feature would fulfill the need of a complementary indicator describing the forecast accuracy and visualizing time-lagged forecast data would further facilitate decisionmaking..

(25) Bibliography [1] National Center for Visual Analytics NCVA url: http://ncva.itn.liu.se.. [cited at p. iii]. [2] L. A. Treinish, ”Visual data fusion for applications of high-resolution numerical weather prediction,” in 2000 IEEE Visualization Conference, 2000, pp. 477480+594. [cited at p. 1] [3] M. Mandiak, P. Shah, Y. Kim and T. Kesavadas, ”Development of an integrated GUI framework for post-disaster data fusion visualization,” in 2005 7th International Conference on Information Fusion, FUSION, 2005, pp. 1131-1137. [cited at p. 1] [4] Swedish Meteorological and Hydrological Institute (SMHI) url: http://www.smhi.se. [cited at p. 1, 3]. [5] European Centre for Medium-Range http://www.ecmwf.int/. [cited at p. 3, 6]. Weather. Forecasts. (ECMWF). url:. [6] W. Mller and H. Schumann, ”Visualization methods for time-dependent data - an overview,” in Proceedings of the 2003 Winter Simulation Conference: Driving Innovation, 2003, pp. 737-745. [cited at p. 4] [7] GeoVista Studio url: http://www.geovistastudio.psu.edu. [8] CommonGIS url: http://www.commongis.de.. [cited at p. 4]. [cited at p. 4]. [9] G. Andrienko and N. Andrienko, ”Visual exploration of the spatial distribution of temporal behaviors,” in 9th International Conference on Information Visualization, iV05, 2005, pp. 799-806. [cited at p. 4] [10] N. Andrienko and G. Andrienko, ”Interactive visual tools to explore spatio-temporal variation,” Proceedings of the Working Conference on Advanced Visual Interfaces AVI 2004, pp. 417-420, 2004. [cited at p. 4] [11] M. Jern and J. Franzn, ””GeoAnalytics” - exploring spatio-temporal and multivariate data,” in Information Visualization 2006, IV06, 2006, pp. 25-31. [cited at p. 4] [12] M. Jern and J. Franzn, ”Integrating InfoVis and geo vis components,” in 11th International Conference Information Visualization, IV 2007, 2007, pp. 511-518. [cited at p. 4]. [13] G. Grinstein, M. Trutschl and U. Cvek, ”High-dimensional visualizations,” Proceedings of the 7 SIGKDD Workshop on Visual Data Mining (KDD-2001)Th, pp. 1-14, 2001. [cited at p. 4] 19.

(26) BIBLIOGRAPHY. 20. [14] J. Blaas, C. P. Botha and F. H. Post, ”Extensions of parallel coordinates for interactive exploration of large multi-timepoint data sets,” IEEE Trans. Visual. Comput. Graphics, vol. 14, pp. 1436-1443, 2008. [cited at p. 4] [15] A. Inselberg, ”The plane with parallel coordinates,” The Visual Computer, vol. 1, no. 4, pp. 69-91, 1985. [cited at p. 4] [16] E. J. Wegman, ”Hyperdimensional data analysis using parallel coordinates,” Journal of the American Statistical Association, vol. 85, no. 411, pp. 664-675, 1990. [cited at p. 4]. [17] ! G. Andrienko, N. Andrienko, ”Construction Parallel Coordinates Plot for Problem Solving,” Smart Graphics ’01 (Pergamon), 2001. [cited at p. 4] [18] P. Lundblad, M. Jern and C. Forsell, ”Voyage analysis applied to geovisual analytics,” in 12th International Conference Information Visualization, IV08, 2008, pp. 381-388. [cited at p. 4, 18] [19] J. C. Roberts, ”Exploratory visualization with multiple linked views,” Exploring Geovisualization, pp. 159-180, 2004. [cited at p. 4] [20] B. Shneiderman, ”Dynamic queries for visual information seeking,” IEEE Software, vol. 11, pp. 70-77, 1994. [cited at p. 4] [21] L. Spirkovska and S. K. Lodha, ”AWE: Aviation weather data visualization environment,” Comput Graphics (Pergamon), vol. 26, pp. 169-191, 2002. [cited at p. 4] [22] Y. Song, J. Ye, N. Svakhine, S. Lasher-Trapp, M. Baldwin and D. S. Ebert, ”An atmospheric visual analysis and exploration system,” IEEE Trans. Visual. Comput. Graphics, vol. 12, pp. 1156-1164, 2006. [cited at p. 4] [23] M. C. Dorneich, O. Olofinboba, S. Pratt, I. Wilson and C. Herbster, ”An experimental evaluation of weather avoidance using route optimization as a decision aid,” in 2002 IEEE International Conference on Systems, Man and Cybernetics, 2002, pp. 608-612. [cited at p. 4] [24] GAV Framework url: http://vita.itn.liu.se/GAV.. [cited at p. 5, 6]. [25] National Center for Environmental Prediction http://www.nco.ncep.noaa.gov/pmb/docs/on388/. [cited at p. 6]. url:. [26] A Guide to WMO GRIB url: http://dss.ucar.edu/docs/formats/grib/gribdoc/. [cited at p. 6]. [27] C. M. D. S. Freitas, P. R. G. Luzzardi, R. A. Cava, M. A. Winckler, M. Pimenta and L. P. Nedel, ”On Evaluating Usability of Information Visualization Techniques,” AVI’02: Proceedings of Advanced Visual Interfaces, 2002. [cited at p. 14] [28] Media Technology and Engineering, Linkping University, Campus Norrkping url: http://www.medieteknik.nu. [cited at p. 14] [29] Lu, C. Yuan, H. Schwartz, B. E. Benjamin, S. G. , ”Short-Range Numerical Weather Prediction Using Time-Lagged Ensembles” Weather and Forecasting, vol. 22; NUMB 3, pages 580-595, 2007 [cited at p. 18].

(27) List of Figures. 1.1. Wind Flags . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 2.1 2.2 2.3 2.4 2.5 2.6 2.7 2.8 2.9 2.10. SWIM overview . . . . . . . . . . . . . . . . . . . . . Weather and voyage visualization in the world map . Change of time step . . . . . . . . . . . . . . . . . . World map level of detail . . . . . . . . . . . . . . . Time step parallel coordinates plot . . . . . . . . . . Ship time graph . . . . . . . . . . . . . . . . . . . . Highlighting in world map . . . . . . . . . . . . . . . Selection area tool . . . . . . . . . . . . . . . . . . . Filtering using time step parallel coordinates plot . . Task flow . . . . . . . . . . . . . . . . . . . . . . . .. 21. . . . . . . . . . .. . . . . . . . . . .. . . . . . . . . . .. . . . . . . . . . .. . . . . . . . . . .. . . . . . . . . . .. . . . . . . . . . .. . . . . . . . . . .. . . . . . . . . . .. . . . . . . . . . .. . . . . . . . . . .. . . . . . . . . . .. . . . . . . . . . .. . . . . . . . . . .. 2 5 8 9 9 10 11 11 12 12 13.

(28)

References

Related documents

Both are central to the ways in which people living with dementia experience and think about the neighbourhoods where they live, and both contribute to a sense of belonging

den svenska identiteten har format utvecklingen och inriktningen av försvaret och det nya svenska totalförsvaret. För att besvara den övergripande forskningsfrågan

In paper I, we have investigated mechanical instability and buckling characterization of vertically aligned single-crystal ZnO nanorods grown on Si, SiC, and

Studien syftar till att bidra med ökad förståelse och kunskap för vilka effekter användandet av CRM har på intern försäljningskontroll.. Den fundamentala och plausibla

Abstract: In this work an Inertial Measurement Unit is used to improve tool position estimates for an ABB IRB 4600 industrial robot, starting from estimates based on motor angle

This research study will use empirical data from earlier presented research papers and research reports done within the Swedish Police (see for example Holgersson, 2005, 2007,

Perhaps because of this, Bury Your Gays appears to have been identified as an undesirable trope. Participants asked about the trope overwhelmingly desired improvement, and

In figure 7 (b) for LHCP incident light, the degree of polarization appreciably decreases at wave- lengths in the band of selective re flection, whereas for RHCP incident light in figure