Transcript Burkholder

Developing a Web Mapping Application using Open Source Software for the Bailey-Matthews Shell Museum
Carrie L. Burkholder
Pennsylvania State University - The College of Earth and Mineral Sciences, Department of Geography
BACKGROUND
1. Bailey-Matthews Shell Museum
a. The Bailey-Matthews Shell Museum is a Non-profit Museum, located on Sanibel
Island, Florida, which is dedicated to the study of Mollusks.
b. The museum features more than 30 exhibits of mollusks from around the
world.
c. Exhibits are devoted to shells in art and history, shell habitat, rare specimens,
fossil shells, focusing on common shells of Southwest Florida.
2. Personal History
a. I got involved with the museum about 18 years ago when I first visited the
Island.
b. My background is biology and I have independently studied mollusks for over
16 years.
c. Interacted and worked with the current and online database.
3. Museum Collections
a. Currently, the Museum encompasses about 150,000 lots of shells, which were
donated or collected from various time periods from all around the world.
b. Museum Volunteers identify and catalogue the shells through an electronic
database built and housed at the Museum
c. In addition, the database is synced to the Museum’s website, which creates an
Online Searchable Collections Database.
METHODS AND MATERIALS
Georeferencing the Database
The database was first standardized and checked for spelling errors or discrepancies and corrected at the request of the
Museum. The database was then split into two new temporary databases, United States and non- United States locations.
The United States database contained 9665 records and the Country database 18179. There were several records that were
removed from the database for several reasons including the site or city was unknown, there was too much information in
the site field which would prevent a geocode match (i.e., 120 km west of Tampa Bay), or the site information was too general
(i.e., Florida Keys, Island, Alligator Alley). Once the process was complete, I had (5246 records) in my USA database. I reran
this same data cleanup for the country data and ended up with (16452 records). I then created two geocode address
locators, one for US Cities with States, the other for World Cities with Country – both had an 80% spelling sensitivity, a 10%
minimum candidate score and a 60% minimum match score. I then ran the Geocode Addresses tool from within ArcInfo on
both datasets. The records that were originally removed as noted above, or those that did not match were hand matched
using Google Maps/Google Earth, the Internet and several printed world maps.
The Purpose of my project consisted of two parts:
1. To create an online web application where a user visiting the Bailey-Matthews Shell
Museum Website can enter an address or location and see what shell specimens have been
found in and near the location.
Part 1 Result:
Currently, the database contains too many records to be passed through the GoogleMap API.
The clustering method seems to work for small numbers of records. I am currently working
with the museum on a sub-set of data, specific to Sanibel Island and Southwest Florida.
Part 2 Result:
The database, populated with latitudes and longitudes was uploaded to the Bailey-Matthews
Shell Museum via ftp site, and uploaded to the internet. The new links (example shown below)
can be found by searching for specific species at http://shellmuseum.org/collection.cfm
Creating the Web Mapping Application
After the database was in a usable format, I began building the web mapping application. The components that I wanted to
include were a geocoding area, which would allow the user to input an address or location and see what shell species were
found and stored in the Bailey-Matthews Shell Museum collections. Another component that I wanted to include was
obviously the map and also a sidebar containing the Museum ID and basic information about the species. The map itself,
along with the sidebar and geocode locator was straight java script and html. I looked at several examples on the internet
and adapted the code to work for me. I initially had my records stored on an SQL server and used PHP to pass to the
records. However, this was extremely slow and did not work for the number of records I was trying to pass to the
application. I then worked with a sub-set of data from Southwest Florida, specific to the Sanibel Island region. Once I
narrowed the records down, I was able to pass the records to the map, however, not all of the records were displaying due to
GoogleMap API limitations. So, I decided to use a cluster method and group the records for multiple occurring records in the
same place. As a user zooms in on the map, more individual records will appear. Custom icons were created for both the
clusters and the individual records. The user can click on the cluster icon to see which species were found in that location.
Once the user zooms in further, small icons appear, which can also be clicked. When the smaller icons are clicked, more
detailed information of the record appear, including the location found, the genus and species and the Shell Museum ID. The
sidebar links are also ‘clickable’, displaying the same information and location as the small icons (Figure 3).
Figure 1. The Bailey-Matthews Shell Museum Exterior and Exhibits.
PURPOSE
RESULTS
Creating the Google Links
Once the database had georeferenced locations for 5% or less of the records, the links to the Google maps were created. To
do this, the records were exported out of the database and imported into Microsoft Excel. From here, I generated a formula
to concatenate the Latitudes and Longitudes of each record. Once the latitudes and longitudes were combined, I was able to
assign other portions of the URL, including the center location, which was the Latitude and Longitude, the type of map,
which was the hybrid map and the default zoom level. Once these records were concatenated, I transferred them back into
the Bailey-Matthews Shell Museum database and saved them as hyperlinks. The museum uploaded all of my changes,
revisions, latitudes and longitudes and has sent them to their webmaster for inclusion in the website (Figure 4.)
2. To assign a latitude an longitude to each shell specimen, allowing the user to search the
existing online collections database and click a link to see the mapped location of
individual specimens.
Figure 4. Part 2 of the project included assigning georeferenced links to the database allowing the user to find a
record in the collections, click on a link and see where the species was found.
CONCLUSIONS
In conclusion, working with the existing database and the records was fairly simple and
straightforward. The most time-intensive part of the project was georeferencing the locations
of existing specimens and overcoming the standardizations, lack of data, sites and spelling
discrepancies within the database. Part one of the project has not yet successfully been
completed for several reasons, including the number of records in the database and the
cooperation of the web master who houses the Bailey-Matthews Shell Museum website.
However, I have been able to create a web application using a small sub-set of records.
Overall, the Museum is extremely happy with the geo-location component which has been
added to their database and each record and have expressed their interest in continued efforts
to create the user front-end web application.
ACKNOWLEDGEMENTS
1. Jose Leal, Director, Bailey-Matthews Shell Museum
2. Ian Turton, Adviser, Pennsylvania State University, Department of Geography
3. Jim Detweiler, Professor, Pennsylvania State University, Department of Geography
Figure 2. The Bailey-Matthews Shell Museum Collections and Volunteers.
Figure 3. Part 1 of the project, the web mapping application, illustrating the clustering effect with genus and species listed on the
left, and the individual records with more detailed record information on the right.