A few words to introduce yourself?
I am Xavier Timbeau, Managing Director at the OFCE. The OFCE is the French Observatory of Economic Conditions, affiliated with Sciences Po. It is a research laboratory for applied economics. The mission of the OFCE - because the OFCE has a mission - is to contribute to public debate on political and economic issues in France and Europe.
Can you tell us a bit more about Debtwatch?
Debtwatch is an online application. It is particularly innovative because we hadn’t done anything like this at the OFCE before, and it’s not something that’s done very often. It’s starting to emerge, but in my opinion, it’s often still a bit embryonic. Here, we aimed for something more ambitious.
The goal was to provide a simulation tool accessible to the widest possible audience, which involves an interface question to access this simulation tool. Moreover, there was also the question about visualizing the results. Simulation results are often complex, so we needed to work on how to visualize the results.
In short: we needed to improve access easier, improve the interface and its usability, and the visualization of results.
The idea behind the simulation is to allow everyone to become an expert on the issue of public debt and, consequently, to participate, discuss, provide information, and share their opinions on these topics.
That’s the ultimate goal. It’s somewhat experimental in this sense, because these aren’t very common practices. Again, there are interactive objects used in communication, but they rarely rely on something very complex behind the scenes. Often, they focus more on data representation with multiples cases.
Here, the ambition is greater, because there’s modeling behind it that works in real-time and in an interactive way.
It’s also more ambitious because it includes many elements that facilitate what we call reproducibility. Reproducibility means being able to run a simulation and allowing anyone to replicate it, diagnose it, and understand the assumptions behind it. This makes it possible to share simulations transparently, because everything used to create it can be replicated by someone else. They can see how the results were achieved, explore what happens if they change one small thing, does it lead to the same conclusion, or does it, on the contrary, reverse the conclusion. It’s about increasing trust in the simulation results..
How was the project born?
To discuss the development of the application, let me go through the timeline. The initial modeling work began about ten years ago. That takes time because you have to to build the model and then use the model in various cases to mature it and bring it to completion. I started creating the first interactive prototypes of this model two or three years ago. It wasn’t using R Shiny but an other software. That’s when I realized it was possible to do things in real-time. Then, I officially started the R Shiny project on June 4, 2021.
It’s true that Shiny allows very rapid prototyping, with cycles of interaction and revision that are also very, very fast. It was not always coded in the most structured way. But in the end, it works.
Why did you reach out to Data Champ’?
The problem is that Shiny isn’t very visually appealing by default. Once I had a certain number of ideas for interaction, I had a number of challenges and difficulties translating them into truly elegant solutions. I was limited by what you can do with standard Shiny usage.
I needed someone to help me go much further, to have something as clean as possible in terms of its interface and its functionality. I also wanted to add some features to the application, some of which I could handle myself. However, it was a lot of work. So I needed to find people to help me and find collaborators who could work on this project within a very short timeframe and with a fairly specific technical scope. Specifically, it was important to stay within R Shiny or deviate as little as possible.
So, I started looking for service providers and then I found Data Champ’.
Why did you choose Data Champ’?
I chose Data Champ’ because their approach seemed more professional than most other providers, which gave me confidence. Additionally, the initial contact and the technical brief were very clear and matched prefectly what I wanted.
How was the collaboration from your point of view?
It went very well. First of all, it was very quick. I gave the costs and it needed to move quickly, with great availability. Your availability was excellent, as I contacted you at the end of July, started the project in early August, and had the work delivered by mid-September. That was pretty quick.
Moreover, I needed someone to integrate into the work I was doing. The idea wasn’t to redo everything but to complement the existing work. It needed to fit together as seamlessly as possible.
When I received the work in mid-September, I had made some progress on my end, so it wasn’t automatic. Initially, I wondered, “Will this work out?” And in fact, it went very well. It took me two days to merge everything because what had been done was well-aligned with the logic of what I had already done. This allowed everything to function without major issues. There were one or two adjustments to make afterward, but they were done very quickly.
What do you think of Data Champ’s work method?
I really appreciated the technical briefs. I found the documents provided, both for the estimate and later for the delivery, were very good because, although I imagine it’s always a bit tedious to do this kind of thing, it’s actually extremely useful for absorbing the work or planning what needs to be done. This method of working is excellent. I believe there will be other applications to develop in the future, and I’m convinced that the model of working with providers on these matters is quite important.
Comments