"National parks are the best idea we ever had. Absolutely American, absolutely democratic, they reflect us at our best rather than our worst." -- Wallace Stegner
While planning our vacations for next year, my wife asks, “What national park should we go to this year?”
“Boy, that is a tough one,” I reply. “Kind of like picking a beer to drink. Lots of things to consider.”
Giving me a skeptical look, she asks, “Are you thinking about using JMP to analyze our vacation plans?”
“What a great idea,” I retort. “I think I will do just that.”
“I can’t wait for the fascinating results,” she says sarcastically.
With my wife’s enthusiasm spurring me on, I get to work pulling in data about the 59 national parks. I pull in the list of national parks from Wikipedia. Then I go the national parks website to get information on visitation over the years and by month. The Wikipedia page has a description of each park, so I take that and run JMP Text Explorer on the description. The resulting word cloud is shown below.
How cool! I can highlight the parks that contain mountains or rivers or any of these words or phrases. I show my wife the word cloud.
“We can highlight the parks that contain any of these words in the description,” I boast.
“Pretty cool,” she admits grudgingly.
“If you are interested in a park with a cave I can right click and select those rows,” I tell her.
“That is nice, but we are planning on a driving trip so we need to know which parks are driving distance,” my wife proclaims.
“Great point,” I admit.
Luckily, the location of each park was in the Wikipedia page, so it was not a problem to add this information. I made a map of each park using that information and with the pictures included with Wikipedia, we could get a preview by hovering over each point. Thinking about my wife’s question, I decide to calculate how far each park is from my house. To do this, I used a formula to calculate millage from latitude and longitude. To get the latitude and longitude of my house, I used Google maps.
sqrt(x * x + y * y) where x = 69.1 * (lat2 - lat1) and y = 69.1 * (lon2 - lon1) * cos(lat1/57.3)
I colored the points by distance from my house. The darkest points are closest to my house in Denver, Colorado. The map is shown below with one of the parks we visited last year (Big Bend) highlighted.
How about a dashboard?
Thinking I might be ready to submit this to my wife again, I debated what else I could do to make this even more exciting for helping select our vacation. I decided to make a dashboard to filter the results so we could narrow down our search even further. I used a dashboard with a hierarchal filter this allows me to filter on one topic then another. The two things used as a filter were the word cloud and distributions on how far the park was from our house. I was finally ready for submission to my wife. I had her inspect the dashboard.
“OK, how does this dashboard work?” she asks.
“First, pick which word or phrase out of the word cloud that you would like to see at the park,” I exclaim excitedly.
“I would like to see a canyon. Now what?”
“Right-click on 'canyon' in the word cloud and choose select rows that will filter only the parks with canyons. Then pick how far you are willing to drive with a 4-year-old,” I tell her.
“How many National Parks can we get to in 10 miles?” she jokes. “Seriously though, we should keep the drive under 500 miles.”
“OK, select all of the bars under 500 in the distance from our house distribution,” I instruct.
She does, and the resulting dashboard is shown below. You can see in the map the six parks that meet our requirements containing a canyon within 500 miles of our house. We can pick between Yellowstone, Black Canyon of the Gunnison, Canyonlands, Bryce Canyon, Zion and the Grand Canyon.
First off, I am amazed that we can drive to six national parks with canyons. Secondly, we are planning on a trip for March, so we wanted to see which parks were less crowded at that time. Looking at a distribution of those six parks, I see that two of them have heavy visitation in March.
If we remove Zion and Grand Canyon, we are down to four parks. Let’s factor in one more thing: the calculation for the average temperature in March. A table of our four options is shown below.
“Well, there is a clear winner here,” says my wife.
Canyonlands, here we come!
I thought it would be nice if anyone could easily do this same analysis of national parks near them. I asked my good friend Jerry Cooper for an uncomplicated way to script this, and he said why not use table variables? He explained you can set these up easily by clicking on the red triangle next to the name of the data table, and add new table variable. The table variable can be used in formulas, and when it is changed, so will the formulas. I put table variables in the attached data table for latitude and longitude. To figure out which national park to visit next, input your latitude and longitude and then run the national parks dashboard script. Here is the new dashboard for Cary, North Carolina (where JMP is headquartered).
Photo credits
- Rocky Mountain National Park photo by Daniel Mayer, posted to Wikipedia, used under Creative Commons license.
- Canyonlands National Park photo by Michael Grindstaff, posted to Wikipedia, used under Creative Commons license.