Using Spatial Analysis Data Visualization Techniques

Introduction

Going out into the community is an aspect of participation for many people.  We are exploring methods for using Global Positioning Systems (GPS) tracking alongside an Ecological Momentary Assessment survey tool (EMA) in a geographic information system (GIS) for visualization and analysis of pain and community participation data. Historically, GPS data have been used for studying transportation and wildlife biology.  Now, using GPS devices for health research is becoming more popular. However, few studies have utilized GPS and GIS technologies alongside EMA.  Using EMA in conjunction with GPS provides an exciting new opportunity to explore the context of health and participation. This paper highlights the preliminary results of such a study and illustrates how GIS and GPS technologies in conjunction with EMA methods can assist with visualizing location data to enrich identification of possible associations between pain and community participation. For this analysis the focus is specifically on pain measures and community participation but the possibilities for additional research are vast.

Sample

The ecological momentary assessment (EMA) participants were selected from the baseline Health and Community survey which queried survey participants about participating in additional research. Of the 525 survey participants, 366 indicated they would participate in future research. Participants were then recruited by phone to participate in EMA training and data collection. A total of 149 participants completed the EMA data collection, 128 of which provided valid GPS coordinates. Of the EMA participants, 99 were between the ages of 18 and 65 years, and 28 were between the ages of 65 and 75, no one who completed the EMA data collection was over the age of 75.

We collected the EMA data using four-inch touch screen devices (i.e., Samsung Galaxy Player 4.0, similar to a smart phone) programmed to prompt participants randomly within two-hour intervals (delivered between 9:00am and 9:00pm), six times daily, for 14 consecutive days.  For more information on this method please see the Pain Interference Patterns Ecological Momentary Assessment research page.

Measures

EMA collected data about an individual’s location, their activities, and pain level. For more information on this method please see the research page at Pain Interference Patterns Ecological Momentary Assessment.

GPS Coordinates

For individuals who provided specific informed consent, the EMA devices collected GPS coordinates every 5 minutes between the hours of 9:00am and 9:00pm. If a GPS coordinate could not be obtained a Wi-Fi access point was recorded instead.  The GPS positioning devices come with a specification on accuracy of 5-40 meters. We do not have an exact measure of the accuracy using WIFI.  All location data was collected in the background and required no input from the respondent.

Participation

Participants were also promoted randomly within two hour intervals between 9:00am and 9:00pm to complete a survey which included questions such as:

  • Where are you?
  • What type of activity are you engaged in?
  • What type of recreation or leisure activity?

Pain

Participant’s also reported their pain level on a scale of 0-10, in response to the prompt, “How much pain are you experiencing right now?”

Methods

Geographers often use tools such as GPS and GIS to investigate human-environmental interactions. This paper explores how time-geography provides a visual analysis of participation, can inform how pain is associated with going into the community.

In order to best visualize the data we explored two visualization techniques.

Technique 1: Tracking Analyst

The first technique utilizes ESRI’s ArcMap’s Tracking Analyst extension to visualize path of travel across days. The travel paths are smoothed and color coded differently for different days (Figure 1).  Blue/green/yellow/orange, and red tracks indicate trips taken during the beginning, the middle, and the end of the data collection period, respectively.

This individual’s map reveals a distinct pattern of activity relative to pain responses, cross the two week period. During the first week of the study this individual made 4 trips away from home (on days 1, 2, 3 and 6). However, only two trips were made across the second week (on days 8 and 9). This second week is shown in the inset.

Figure 1. Visualization of out-of-home travel paths for the individual during data collection period with ESRI’s ArcMap Tracking Analysis Tool.

Travel Paths

 

 

Figure 2. Reported pain levels of the individual during data collection period.

Pain Levels

 

 

Matching this spatial data to the survey responses in the EMA allows us to search for any possible relationships between participation out in the community (trips away from home) and reported pain levels. Figure 2 shows this same individual's pain responses across the two week data collection period with colors in the chart corresponding to the colors in the map.  For the first week of data collection this individual's pain level fluctuates between a 1 and a 4. However, on day 8, the pain rating jumps from a 1 to an 8 and remains high for the rest of the week. This pain spike is shown in yellow on the map and the graph. The jump in pain level seems to correspond to the trips away from home. On the day of the pain spike the individual takes a trip, as well as the following day, however, for the remainder of the collection period, when pain is still much higher than the previous week, this individual remains at home. Viewed in conjunction with the graph of pain level rating across time this map is useful for visualizing how one individual’s trips outside of home are related to their experience of pain.

This analysis is limited in a couple of ways. First, this map effectively shows path of travel by day across a two week period but reveals little detail about travel within days. For example, it is difficult to visualize how long an individual spent a different locations, were they just passing through or was it an extended stay? Did the trip last all day long or just a couple hours in the morning? These details cannot be easily displayed using the tracking analysis.

Visualization 2: Space-time paths

The second visualization, space-time path visualization, is able to address some of the limitations of the first. The visualization technique used here is based on Hägerstrand’s concept of space time path as an illustration of how a person navigates his or her way through the spatial-temporal environment. This is shown in Figure 3. The space-time paths represent the individual’s trajectory of movement over space and time in a 3D visualizing environment. The vertical axis represents time, while the horizontal plane represents the geographical extent of the individual's activity. If an individual stays stationary, the path will appear as vertical. If the participant moves, a line is drawn connecting the movement origin and destination of travel. The slop of the line represents the speed of movement. For days involving no community participation, their space time paths are vertical lines rooted at their home locations.

We used Visual Basic for Applications (VBA) programming that used ESRI’s ArcObjects to create space-time path features. These features are then visualized using ESRI’s Arc Sense application.

Figure 3: Space-time paths

Hägerstrand’s space-time path

 

 

Space-time paths corresponding to the time stamped and survey linked location data collected for participants in this project are visualized for daily intervals. Figure 4 depicts daily space-time paths of the same individual whose travel paths are shown in Figure 2. The temporal range for each path is 12 hours, from 9 AM – 9 PM. A color ramp from blue to red is used to distinguish space-time paths of day 1 to day 14 of the data collection period.

Reported pain levels by the individual during the data collection period are also mapped onto the space time path using a consistent color ramp with space time paths to indicate pain level for different days. Small to large proportional symbols corresponding to pain levels from 0 (low) to 9 (high) are used.

Figure 4: Visualization of Daily Space-time Paths for the Individual During Data Collection PeriodIntermittent Pain image

 

Upon examination of the map, it is clear that a majority of high pain reporting happened at home and was recorded during the 2nd week of data collection. Finally, like the previous map, this illustrates how little this individual went out into the community after a couple days of high pain activity. The yellow to red shaded lines represent movement during the second week of data collection. For this individual, there are no red tracks leaving home, only blue and beige, indicating that their community participation took place primarily during the first week of the study.

Conclusion

This study and brief analysis reveals that the combination of EMA and GPS data collection techniques presents exciting opportunities in the intersections of geography, disability and health research. Dynamic real-time measures provide insight into the complex relationships between pain, disability and participation. These visualizations offer a new and exciting method for illustrating health and participation data and may offer more opportunities for analysis. Further exploration is needed, potentially utilizing large datasets to identify patterns and relationships across individuals and subgroups on a larger scale.