Prototyping for digital rights lobbying
Raising the profile of digital rights as human rights at the United Nations. One line of code at a time.
Do you ever replay in your head a conversation you had recently? Only this time, you have that perfect reply ready. Or an insightful comment. Or a fascinating fact.
What if that conversation was about digital rights in your country? And you were talking to a UN diplomat who actually has the power to do something about it? You wouldn’t want to miss your only opportunity to make a difference.
That’s what we’re trying to build at DATA4CHANGE. A tool that helps civil society organisations (CSOs) lobby the countries making recommendations during the Universal Periodic Review (UPR) more efficiently.
Not “the-perfect-comeback” machine exactly. More like the “get-your-facts-ready” and “get-your-point-across” machines combined together. And we’re using AI to do that!
Well, not really. Sorry. I just wanted to see you getting all excited for a moment. We’re using Python.
The task at hand
So yeah, we want to help CSOs find the right people to talk to, and then teach them how to talk with them… Oh, and prep them for that important meeting. With the very important people in Geneva. That’s right. In Switzerland.
We have a lot of expert knowledge inside DATA4CHANGE. We also have 60,000+ previously made UPR recommendations.
Ideally, we want to create a product that uses available data to replicate our expertise. A product that allows people to become their own expert. Without us needing to be everywhere. All the time.
60,000 rows of data is a lot. Reading it would take you like at least three days. Probably. You do the math.
To work with all that text data in a sane way, we need tools. There are some available already. Excel for example. It’s great. Not for that amount of data, though. We’ve tried that, didn’t work.
With so many to choose from, how do you decide on the right one for the job? We decided to go straight for code and prototype in a browser. You want to know why?
Good. I’m glad you’re still here.
Prototyping in code enables us to do *anything* you can do with the modern web. We probably won’t have to. But it’s important to know we can, because ultimately we’re building a web application. And staying in our target medium all the time helps us keep all the affordances in mind.
Prototyping is an important part of our process. It helps us figure out what we’re doing. And have a conversation. It allows us to take *the* greatest idea, implement it real quick, and then be like “yeah, this isn’t really working, is it?”
And it enables us to find the best possible way for the task given. Quicker. To get the most out of everyone’s time.
So far we’ve spent a month exploring different ways of looking at our dataset. We wanted to show you the best stuff. Also, the stuff that’s the most relevant to you.
Actually, we just wanted to show you what you’re looking for. Even though you might not know what that is.
Yeah, you’re not easy.
The best stuff
We start by showing an overview of the countries making the most recommendations. And on which topics recommendations are made most often. If, like everyone else, you’ve got limited time and resources to do advocacy, then you’re going to want to know which countries are the best to target for your particular topic.
Want to get a feel for the dataset? Not sure what you’re after? This is your gateway.
The stuff you’re looking for
Once you know what you’re looking for, you can start digging. And keep digging. Until you find that one detail you really need. Until you find how many recommendations Czechia made to Jordan on the topic of Torture in the first UPR cycle.
That’s very niche. Which is good. You’re special. And you might be only interested in one or two recommendations.
Whatever you feel like at the moment
To allow you to jump seamlessly from one topic to another, we’ve also added a search box where you can type the ‘whatever’ it is you and you alone are looking for. A country, a topic or a keyword.
This is the freestyle mode. Simply start typing and select the best matching result.
Ideally, we would like the whole thing to be just a search input. Ask a question, get an answer. Maybe even a voice assistant? “Hey UPROAR, who should I talk to about digital rights in Jordan?” Well, we’re not there yet. But we will be.
Who to talk to
To help you find the right recommending country to talk to, we are designing a Friends&Foes™ index. It takes into account the history of recommendations made and received between pairs of countries. And some other variables (aka secret sauce).
We then take those number, move them around a bit and give you a single score.
Or a smiley/frowny face. Or maybe a red/green light? We don’t know yet. What we do know is that we want you not to waste your time talking to the wrong people.
Looking at the list, you’d probably want to hedge your bets and set up a meeting with France to make a nice friendly recommendation that’s likely to be accepted.
And also a meeting with Norway to get a hard-hitting one on the table as well.
How to talk to them
Once you know who you should talk to, it’d be handy to know what kind of recommendation you would like them to put forward.
If you read some of the UPR recommendations, you quickly realise that countries have funny ways of communicating with each other. Each recommendation usually starts with a verb. And that verb tells you a lot about how strong the recommendation is.
Putting all the verbs together helps you understand quite a lot about what the diplomatic (or not so diplomatic) relationship is like between countries.
Understanding the nature of the relationship between countries is important. You want to approach the country you’re going to ask to make a recommendation with the right language. Be diplomatic when appropriate, forceful when necessary.
We’re no futurist but we think unlocking the key to these verbs will be the defining feature of our tool. We’ll use it to advise you how to approach different countries. So that you can talk in their language.
We want to build a sophisticated product which is extremely easy to use. You know. Like every other team building a data-driven product.
But ultimately we want you to find the right people and impress them. Make them weak in their knees. Make them go: “Great job team, we’ll be in touch.” And have them make YOUR recommendations, verbatim, during a Universal Periodic Review session.
We want to make your job easier. So that you can make their job easy. Win-win.
We are right at the beginning, with lots ahead of us. We promise to work hard. So stay tuned.
Maybe you’ll even get your AI.