upa - home page JUS - Journal of usability studies
An international peer-reviewed journal

Development and Evaluation of Two Prototypes for Providing Weather Map Data to Blind Users Through Sonification

Jonathan Lazar, Suranjan Chakraborty, Dustin Carroll, Robert Weir, Bryan Sizemore, and Haley Henderson

Journal of Usability Studies, Volume 8, Issue 4, August 2013, pp. 93 - 110

Article Contents

Design of the First Sonified Weather Map Prototype

In the following sections we describe the details of the system architecture and user interaction to our first prototype.

System Architecture for Retrieving Weather Data

We built the accessible weather map by modifying iSonic. The iSonic application in its current version works with population data. It is designed to work with only static data and requires the source data to be in a CSV file format. While population data is updated maybe once a year, weather data is updated frequently, often hourly. For the proposed weather map sonification, we needed to modify the application to work with a real-time data feed.

The source for the weather data was the National Oceanic and Atmospheric Administration (NOAA) website, which is run by the U.S. Federal Government and therefore provides weather data to the public at no charge and with no permissions required. The NOAA website provides real-time weather data in raw textual format that met the information specification of the proposed weather application and was formatted for easy retrieval. The data querying was developed using Microsoft Excel macros that allow retrieval of external data from tables located on HTML pages. The weather data for the different counties within Maryland was retrieved through the manipulation of the longitude and latitude values in the data retrieval queries. This allowed the importing of weather data for multiple points in Maryland into an Excel workbook. The particular data points were developed through the compilation of a list of cities that were completely within the boundaries of each county, usually the county’s administrative center. We assumed that these cities would be representative of the weather within the county in most instances because most of these counties are not geographically very large.

We implemented the data feed from the NOAA web server to the iSonic map interface using the following steps:

  1. Developed an Excel sheet (described above) to import data from the NOAA web server.
  2. Created a second Excel workbook and formatted it appropriately to make it readable by the iSonic application.
  3. Linked the cells of the two Excel sheets so that the second Excel sheet could automatically pull data from the Excel sheet containing imported weather data from NOAA’s web page.
  4. Developed a java program allowing automatic storing of the data from the second Excel sheet as a CSV file. In addition, we created VBA macros for both Excel sheets that initiate a refresh, save, close window sequence.
  5. Created a batch file so that the weather data was automatically updated and converted into a file readable by the application.

Figure 2 shows the system architecture schematic for the data retrieval process. In the next section we provide descriptive details about user interaction with the iSonic weather map application.

Figure 2

Figure 2. System architecture schematic

User Interaction With the iSonic Weather Map

We developed the prototype iSonic weather map application by overlaying weather data on the native iSonic application. A consequence of using the native iSonic platform was that we were constrained to use the existing mapping between data and auditory feedback in this platform. This was clearly a trade-off. The mapping between data and auditory feedback was based on static population data, not frequently changing weather data. It is possible that another type of mapping would, from a usability point of view, have been superior for weather data. However, given the lack of funding for the project and the limited schedule, we decided that the prototype would be based on the existing iSonic application. While this represented a design dilemma, we decided to accept this as a constraint for this exploratory study and also to use this configuration as an opportunity to test the usefulness of the existing mapping.

We developed the first prototype weather map to provide a user with three different types of weather information—temperature, wind speed, and percentage chance of precipitation (see Figure 3 for the temperature example). We made a design decision to present a user with each of these types of weather related information individually (only one type of data per weather map screen) to reduce information clutter (this decision was based on the feedback received during the requirements gathering).Therefore, we designed the prototype weather application to allow a user to switch between three separate sonified maps that provided weather related information on temperature, wind speed, and percentage chance of precipitation, respectively. Figure 3 below provides a snapshot of the sonified Maryland map.

Figure 3

Figure 3. Prototype map interface (The counties with darker shades of green have higher temperatures than the counties with lighter shades of green.)

As mentioned before, the iSonic native application interface provided the environment for user interaction. The primary interaction mechanism in this interface is keyboard based. For the prototype and usability study, an additional interaction strategy was adopted. This involved using a touchscreen with a tactile map overlay to traverse through the iSonic weather maps. The tactile map had raised edges for the contours of map boundaries. The raised edges represented county borders within the state of Maryland. The tactile map did not have any braille. While the original iSonic application had the capability to use a touchscreen for interaction, the touchscreen capability had not been previously evaluated for usability (Zhao, Shneiderman, Plaisant, & Lazar, 2008). For this study, a KEYTEC Magic Touch touchscreen that was calibrated to the iSonic map application was used. The initial plan was to use a tactile map of Maryland manufactured by the National Federation of the Blind’s printing facilities. However, that tactile map did not calibrate to the scale expected by the touchscreen. To solve this issue, we created a tactile map by hand. We placed a piece of overlay paper on the touchscreen that displayed the iSonic map. We used a pencil to trace the map of Maryland from the iSonic interface to make a visual representation of the map on paper. Using needles (not near the touchscreen), our team “poked through” the paper and created tactile contours of the map boundaries using needles so that the map boundaries could be ascertained using tactile means. One of the challenges of doing so was that the overlay paper couldn’t just be any paper, but it needed to be paper that was thick enough to not rip but thin enough that the touchscreen could still sense human touch. The right “level” of paper, usable for tactile maps over touchscreens, was provided by the National Federation of the Blind.

The iSonic interface provides two information presentation choices (or data views) to a user. The first is the default map-based representation. The second is a tabular data view where the first column represents the geographical elements of the data (counties of Maryland in this instance), and the other columns represent the specific data domains being presented (e.g., temperature in Figure 4). Each row in this table represents the temperature, wind speed, and chance of precipitation for a Maryland county within an hour of the last data collection from the NOAA website. A user can switch between the two data views using the TAB key. The tabular data view allows a user to toggle between the three different weather maps (for temperature, wind speed, and chance of precipitation). For example, users can move from the temperature map to the wind speed map by pressing the TAB key. Users can switch from the map to the tabular view by pressing the TAB key too, and then press the arrow key to change the data focus to the wind speed column, and then press TAB again to move back to the map. Figure 4 shows a view of the tabular view showing temperature data for the Maryland counties.

Figure 4

Figure 4.Table view of the iSonic weather map

A user has two different choices of navigation through the map—absolute and relative. During both absolute and relative navigation, the map is traversed in units defined by the geographical boundaries of the Maryland counties. The absolute navigation, which is started with a simple key press (once started, the entire map is swept), allows a user to sweep through the map to get an overview of the values. This sweep is done left to right and from top to bottom. It is useful if a user wants a quick overview of the weather across Maryland. For example, if a user wanted to get an idea of precipitation across the state he/she would need to choose the precipitation map and initiate the absolute navigation. The iSonic application will start a sweep of the Maryland state, west to east and then north to south. As the sweep is made, a user will get feedback about the chance of precipitation for all the counties within the state. A user will also hear a percussion sound at end of each row of the sweep and a bell sound at the end of the sweep. Alternatively a user can perform a more controlled and detailed exploration of Maryland by using the relative navigation. For relative navigation, a user can press the four arrow keys to navigate up, down, left, and right in the map. By default the relative navigation commences from the left, top corner of the map. A chirping sound alerts a user when the navigation takes him/her outside the boundaries of the map (e.g., outside the boundaries of Maryland in this instance). As a user traverses through the state, they are provided auditory feedback about the weather-related data for that county. The relative navigation also allows a user to focus on a particular point of the map. For example if a user is interested in the weather in Baltimore County, they could traverse to the county and then get the specific weather details (temperature, wind speed, pressure, and chance of precipitation).

A user has the choice of receiving this feedback with different degrees of detail or information level. The least detailed is level 0, providing the name of the region (in this case, the county), but no data. Data level 1 provides a tonal sound to represent the data (e.g., the temperature), but no additional information. The pitch of this sound is used to represent values of the data, a higher pitch being associated with the larger value (e.g., high pitch for high temperature and low pitch for low temperature). Level 2 augments this tonal sound with speech output, such as the name of the county. Level 3 includes the pitch, the county name, and the temperature (or wind speed or precipitation) in the county. The level of information in the feedback can be increased or decreased by pressing the “+” key or by selecting the information level choice in the menu bar. In Table 2 we provide the details for the different levels.

Table 2. Feedback Levels

Table 2

Therefore, users can navigate to a particular county and choose the level of detail they would like. For example proceeding with our earlier example, if users interested in weather in Baltimore County chose the temperature map and increase the feedback level to the detailed level, they would receive feedback that would inform them of the name of the county, the sound pitch corresponding to the temperature, value of the temperature (e.g., 32 F), and the fact that the specific weather attribute being referred to is temperature. On the other hand, if users wanted to explore the Maryland geography, they can choose level 0 and traverse the map using absolute or relative navigation. Similarly, users interested in temperature trends across the state could choose the temperature map and initiate absolute or relative navigation. The increase and decrease of the pitch as they traverse the map would give them a sense of the trend of temperature distribution in the state.

For this study, participants also used a touchscreen with a tactile map overlay to navigate the Maryland map. With this alternative, participants selected a particular county by tracing its contours on the tactile map and then tapped on it to get weather related auditory feedback for that county.


Previous | Next