ed later, analyzes the coordinates and renders an icon of palanquin on the digital map according to the coordinates. Real-time videos captured by the DVRs setup at four corners of the palanquin are stored in the hard disk. Then, 16 frames within a 2-minute interval of each video are selected and sent to the control center via GPRS module every 2 minutes. Figure 3 is the photo showing the installment of hardware on the back of Mazu’s palanquin described above.
The lower part of Figure 2 summarizes a collection of three servers responsible for different software services. The WebGIS server is the Microsoft web server, or Internet Information Server (IIS), where the portal embedded with a web-based map platform is located. The map platform is EasyMap, a product developed by GIS Research Center, Feng Chia University using ASP.NET web programming, which provides map data and application programming interfaces (API) to manipulate and display map features.
Figure 2. System architecture.
Figure 3. Hardware installment.
DVR server is responsible for managing video data sent by those four DVRs. The selected video frames are then embedded in the portal as well as in mobile portal, which is prepared by Mobile GIS Server separated from WebGIS server to down size of system loading.
Since 2011, the development team offered both i-Phone App and Android App for the public to download and access the mobile portal by smartphones. Pilgrims and tourists who have smartphones can download these free applications from Apple’s App Store and Google Android Market, respectively. These smartphone applications can display the palanquin’s location in Google Map, instead of EasyMap, and can be used by PDAs and desktop web browsers. It also provides historic trails of Patrol and Pilgrimage on previous days as well as the real-time video with high resolutions. With these applications on mobile clients published to the public in 2011, the number of visits to our services dramatically reached up the summit since 2008, which also reveals that mobile devices could play an important role in diffusion of dailylife technologies.
From several informal interviews to pilgrims and tourists using our mobile clients, we learned that our services could provide not only the guidance of pilgrimage and patrol, but also a guidebook of tourism in those areas along the route. Staff from Jenn Lann Temple also used the mobile clients for coordination, where the accurate position of palanquin helped with oral communication as well as estimation of time left to prepare for some measurements before the palanquin arrived.
4. Statistical Analysis
Data transmitted by the GPS car kit every 30 seconds contain latitude and longitude coordinates and transmission time. The accumulation of data in the database is rapid and tremendous. For example in 2010, there were 11,526 valid records stored in the database. However, those spatial-temporal data can be an important source for statistical analysis, and thus reveal some cultural meanings behind as well as provide a reference for festival organizers to make decisions. This section presents several statistics collected and analyzed from the past three years, including patrol and pilgrimage route, mileage, idling time, and discusses some important issues from cultural and administrative perspectives.
4.1. Patrol and Pilgrimage Route
According to literature, the planning of patrol and pilgrimage route depends on several key factors, such as the serving temples, traffic concerns, and non-repetition [5,7, 12]. As mentioned earlier, purpose of the patrol is for Mazu to inspect the area under her control; thus those temples within the area naturally taken into account of patrol route. However, temples that participate in the pilgrimage may vary from year to year. For example, in 2010, Jenn Lann Temple decided to expand the pilgrimage, and threw “divination blocks” to ask for Mazu’s permission for additional day of pilgrimage, which resulted in as many as 27 temples to participate in the festival for their very first time.
Cross-country routes, whose main consideration is the convenience of traffic and accommodation, are generally along with provincial road number 1 of Taiwan, which is wide and roughly straight, while routes within the township adhere to the principle of non-repetition, in order for palanquin to meet the demand for more local believers. In addition, the forward pilgrimage is subject to arrive in the destination for the scheduled ceremony, so the route on forward trip within the same area often shorter than the backward trip. Figure 4 is the route of patrol and pilgrimage where red line indicates the forward trip, much shorter than green backward trip. Figures 5 and 6, show both forward and back routes in both Changhua and Yunlin County in 2010, respectively, indicating the principle of non-repetition within a township.
Figure 4. Entire patrol & pilgrimage routes in 2010.
Figure 5. Pilgrimage routes in Changhua County.
Figure 6. Pilgrimage routes in Yunlin County.
4.2. Walking Distance
Total walking mileages of patrol and pilgrimage from 2008 to 2010 can be obtained by linking consecutive positions of the palanquin, as show in Figure 7. The total walking mileages roughly range from 295 to 300 km, where the forward trips by and larger are 15 to 30 km longer than the backward trips. That’s due to the forward trip subject to a fixed arrival time for scheduled ceremonies as mentioned earlier .
In 2010 when 27 more temples joined in the pilgrimage and one more day was added, however, the total walking distance didn’t significantly increase; instead the palanquin spent more time on temples along the route and slowed down the speed on the backward trip. This arrangement can be found in Figure 8 where daily walking distances for three years are plotted.
Figure 8 also reveals a fact that the forward trip is one or two days less than the backward trip. The forward trips are always 3 days long while the backward trips are 4 days in 2008 and 2009, but 5 days in 2010. The longest daily walking distance falls in the second day of the forward trip for three years, with a mileage up to 58 km within a walking period of 18 hours. That’s because on the second of patrol the palanquin always passes Chunghwa
Figure 7. Total mileages of pilgrimage from 2008 to 2010.
Figure 8. Daily mileages of pilgrimage from 2008 to 2010.
County where around 40 temples join in the patrol, accounting for about one third of the total participating temples. The palanquin doesn’t move on day 4 because of ceremonies, and thus the second longest walking distance falls in the first day of backward trip, or day 5 for the entire journey probably because the pilgrims gain more energy from resting on day 4.
4.3. Idling Time
Factors of affecting the idling time of palanquin at a certain temple includes the fame and scale of temple, the relations between the temple and Jenn Lann Temple, the enthusiasm of the people along the way, and “Robbing Mazu’s palanquin”.
Idling time can be divided into three clusters with different length groups. The first cluster of idling time is 2 to 3 hours, followed by the second cluster of 30 minutes and the third, less than 20 minutes. Temples with long idling time are those sites with high potential to become spots of tourism.
4.4. Website Analytics
Estimates of pilgrim and tourist number by the press, governments, or Jenn Lann Temple could range from hundreds of thousands to millions depending on the methods of estimation. However, the trend of increase in popularity of the event since its early years is obvious. The promotion by Jenn Lann Temple in recent years is generally acknowledged as the major driving force to this tendency. Adoption of information technologies is one of measurements used by the temple authority.
Since 2008 when the reporting services of Mazu’s palanquin was presented to the public for the first time, the development team has adopted Google Analytics, as shown in Figure 9, as the tool of monitoring the website traffic to filter malicious attacks as well as to analyze the geographic distribution of portal visitors.
According to summaries of Google Analytics on our portal, the average year-by-year increase rate of portal visitor numbers during the festival from 2008 to 2011 is 16.46%. With the provision of smartphone applications in 2011, however, it dramatically increased by 33.77% compared with the year of 2010. Figure 10 shows the distribution of mobile devices used by visitors to access the mobile portal. Android had almost twice as many users as iPhone among our portal visitors. Both major mobile devices accounted for about 85% of visits.
With those analytics being monitored every year, we can trace the portal visitors, where they are from or how long they stay on the site, and thus estimate the growth of event popularity or even the level of culture dissemination. Figure 11 illustrates the geographic distribution of visitors by cities around the globe in 2011. Over 180 thou-
Figure 9. Website Traffic of portal by Google Analytics.
Figure 10. Distribution of visitors by mobile devices.
Figure 11. Geographic distribution of visitors by cities.
sands visits came from 200 different cities, mostly in Asia, North America, and Europe.
4.5. Error Analysis
GPS are still subject to satellites errors, tropospheric delay, ionospheric delay and other factors, and therefore its accuracy is merely 10 to 25 meters. In addition, GPS signals are highly affected by vigorous vibrations, especially when believers set off firecrackers while the palanquin passes by. The vibration of the palanquin would often result in abnormal reporting of coordinates by the GPS receiver (see Figure 12).
Figure 12. Abnormal GPS coordinates.
These abnormal coordinates need to be removed manually (see figure 13 for the outliers with a value higher than 200 meters). Therefore, strengthening the stability of hardware and developing an algorithm to automatically filter erroneous coordinates, such as to reduce the manual work and increase timeliness and reliability of information.
This article describes how to apply spatial technologies to Dajia Mazu Patrol and Pilgrimage, including the positioning
Figure 13. Analysis of abnormal coordinates.
technology, equipment and operation framework adopted by the initiative. It also explores how spatial technologies can help organize the large-scale event. Publishing the real-time position of Mazu’s palanquin on the internet can prevent pilgrims and tourists from tortures of long waiting. The historic trajectories can also assist festival organizers in choice of pilgrimage routes, and re-examine the planning in hindsight to improve future festival . The following paragraphs summarize conclusions and future studies for public reference.
● This study successfully applied spatial technology and wireless communication to publish real-time positions of Mazu’s palanquin and live video of pilgrimage on the internet, facilitating the festival proceeding. We demonstrated that technologies can boost the tourism boom, which could be a fusion paradigm among culture, tourism, and technology.
● During the patrol and pilgrimage, both battery of car kit and memory card of DVRs need to be replaced once every 12 hours. The system should solve the battery and memory card problems to reduce replacement frequency of equipment and thus the number of data loss.
The current system refreshes the video once 20 seconds with JPG images of 180 × 131 pixels. There are still significant deficiencies on image’s clarity to be improved in the future.