On May 17th, a small team of researchers and Ba'Yei polers set out to explore the Okavango Delta's catchment from top to bottom. By the time the expedition team pulls their boats up into the sands of the Kalahari at the end of August, they will have paddled, pushed, and poled more than 2,000km over 100 days since their launch in the misty Angolan highlands. They will have eaten more than a thousand kilograms of beans, swatted a small army of mosquitos, and threaded their narrow mekoro through dozens of pods of angry hippopotami. They also will have collected a lot of data, all of which is freely available through an open API:
7,777 wildlife sightings, 17,822 sensor readings, 2,042 images, 105,566 GPS points, and 3.1 million heart beats so far.
When I started working on the Into the Okavango project three years ago (it's a nine year effort), I believed that making the expedition data open would be enough. In my mind I concocted a fantasy audience, a group of JSON-savvy, conservation-minded enthusiasts just waiting to spend their evenings and weekends making fascinating things with our datasets. This is the open data dream. But with our first attempts, most of these interested parties (if they existed) were being blocked at the door, sometimes by technical barriers, but more importantly because the data didn't offer contextual and narrative hooks that they could latch onto.
Open datasets might be fascinating, the APIs might be usable; but without any structure for story, it's hard for anyone to make anything with them. Perhaps because of this, public data releases too often end up like teetering piles of free IKEA parts, kicked out to the curb with no instructions.
I've learned a lot during my work on ItO over the last three years: how to stand down a charging elephant, how to walk with cracked heals, how to say good morning in Tswana. Most imortantly, though, I've come to believe that open data efforts can only be most effective if they are paired with open story.
The sound you're hearing in the background (double check that your speakers are on) is a sonification of expedition leader Steve Boyes' heartbeat from Saturday, July 11th, at around 10:58am. He's just about to be attacked by a three tonne hippopotamus:
"The front tusks pierced the hull near my left foot and flipped the mokoro over in a single push. Capsized, we were now both in the water on top of the hippo. My first thought was get my legs out of the water. I met Giles on top of the overturned hull. I hear my brother, Chris, shout.
"Swim!!"
We instantly slip into the water and make for the bank, frantically swimming like no one taught us how. I have never stressed, raced and imagined so much. Seconds of frenzy that I will never forget. We were convinced the follow-up rush and bite was coming. Giles touches my leg and my heart stops beating. The current was taking us and the bank seems further away. Suddenly it was there. Slippery and steep."
This particular hippopotamus lives in our database as a GeoJSON record. Also available are about 3,000 numbers, intervals between heart beats in a 20 minute span, as measured by a chest-strap heart rate monitor that Steve wears every day while on expedition. The graph below shows all of Steve's thirty-five thousand heartbeats from July 11th. If you click and drag on it to select the bright pink stripe of data, beside the white line, you'll be able to see the harrowing cardiac sequence more closely. You'll also be able to move your mouse to compare the sound of the peak heart rate with some of the calmer moments before and after. Go ahead and do it.
This visualization of the data shows us a lot about that day: the duration, the pattern of activity, the maxima and minima. The sonification, concurrently, offers us empathy; when Steve's heartrate accelerates, ours does too. This reaction to an audible heart beat is instinctive, a leftover from our time in the womb. Finally, the written narrative brings the experience into the first person, allowing us to imagine what it might be like to be in Steve's sandals.
With our efforts on Into the Okavango, we've set out to enable a series of expeditions that are both Open Data and Open Story. We want to build a system that not only shares the scientific results of the expedition, but also provides an accessible architecture for building narrative on. Here are our three guiding principles :
Wherever possible, we've endeavored to provide framing for the data that we release. For a simple example of this we can look at the visualization and map views in our API. Using small variations of the same URL, people can download sightings as raw data, or see the results on a map, or as a graph:
I often describe APIs as bridges. I think one of the issues we have right now is that these bridges are being built for heavy machinery, and not for foot traffic. What does an API look like that is a pedestrian bridge?
We want our API to be usable not only by researchers, but also by students, teachers, journalists, artists, musicians, and the generally curious. In service of this, we've tried to provide simple access points to data that anyone can use. We're also releasing samples and source code on GitHub, and are working to create a series of curricula for use by students and teachers.
Though we have a long way to go, results have been promising. Students in National Geographic education programs have been using our GPS points to learn about the geography of Western Africa. Shah Selbe has led workshops for youth, in which they build open source sensors based on the ones that we've been deploying in the field.
Science doesn't just happen. Humans do science. Humans with real lives, real foibles, real blistered feet. As much as possible, we've tried to pair data about human experience (tweets, photographs, sound recordings, Medium posts) alongside the core research data. And we've kept this human data in the same format as the science data (GeoJSON) so that it's easy to thread one data set through another.
On our website, people can see the progress of the expedition in terms of individuals and their exact paths, and can see the scientific data in conjunction with human-centered content:
#Okavango15 ends in a few weeks, but we still have four more years of transects to do, and many millions of data points to share. I hope that as the mission continues, we'll be able to build on this Open Story foundation that we've built. The end goal for the project is to drive meaningful conservation efforts in Angola, Namibia and Botswana, and we'll only be able to do this with the combined support of the scientific community and of the general public. We'll need to be able to tell our story and the story of the Delta alongside the data that we release. For this we'll have to support storytelling alongside analysis, and we'll need to find ways to foster empathy as well as understanding.
As you were looking at Steve's heart rate graph you might have noticed a second spike, about an hour after the hippo attack. This confused me; from the GPS points I knew they had continued down the river after only a short break. Had there been another scare? As I was drafting this post, I sent him an e-mail, asking what he thought could have caused that second elevation in heart rate. A few days passed.
Finally, an e-mail arrived. Steve told me that a few short minutes after the attack they got back into the boats. He tried to stay as calm as possible, as he paddled down the wide river. After forty minutes he steered his boat to the bank, and anchored it loosely in the reeds. He took out the satellite phone and called his wife.
He told her the story you just read, about the tusks piercing the hull, about dangling helpless in the river, about the frantic swim to shore.
You can follow the #Okavango15 expedition as they transect the heart of the Delta on Twitter at @intotheokanango and on Instagram.
Jer Thorp is an artist and educator living in Brooklyn. He is a National Geographic Emerging Explorer and a co-founder of The Office for Creative Research.
I'm a fan of Medium. Not only do I publish a fair number of articles there, I also use it as a word processor whenever I have to write pretty much anything. My intention with this article is not to rip the Medium style, but to suggest how data elements could be added to its already excellent design for reading.
Many, many people have helped in the creation of the ItO API + site. Special thanks to Brian House, Shah Selbe, Ellery Royston, Ian Ardouin-Fumat and Genevieve Hoffman
Visit intotheokavango.org to explore all of the data that has come in from the field since the beginning of the expedition. Also, follow us on Twitter & Instagram for a curious mix of wildlife and data.
If you're specifically interested in heartrate data, here is the API call that this example is using, which returns all of Steven's heart beats from day 68 of the expedition: http://intotheokavango.org/api/features/?FeatureType=ambit_hr&expeditionDay=68&Member=Steve
All of the interactive bits here were coded in p5.js, which is my new favourite thing. This whole article is available in all of its source-codey glory on GitHub.