I had been holding back on writing this as a blog series because I was toying with the idea of writing an MSDN Magazine article about it – but honestly, I think that MSDN Magazine probably has more important things to write about than my pet project. That said, this pet project is one that I’m pretty happy with, so I’ll share it with you here.
Several months ago, I came across Pivot, a project that was back then just another a research effort coming out of Microsoft’s Live Labs group – but it was a research project that had recently received some notoriety as a result of Gary Flake’s TED Talk. After watching the talk, I was inspired to play around with the technology – so I downloaded the thick client viewer, ran through the samples and voila – I had my first collection. Put another way, I now had a really cool solution without a problem.
So what type of data might be a good real-world fit for visualization in Pivot? Well, I had a pretty decent understanding of MSDN Magazine data, so I hopped over the Web site to take a look…
Seriously, is there any way this could be screaming louder for a rich, pivot-style UX??
From looking at how the site is structured today, it seems clear that at some point, we’ve made an assumption that people wanted to be able to look at content across a variety of different dimensions. However, when you look at the page traffic data, you can see that those pages (topics, authors – which we in fact took down because it was hardly ever viewed, columns, etc..) are not viewed all that often. There are a couple different ways to explain this – 1) people simply don’t have the patience (nor should they) to wade through hundreds or thousands of links in a list, or 2) people simply don’t care about exploring content and would rather just go back to search. I want to believe that the answer is closer to #1 – that effectively, the problem is that we simply have a site that is not compelling for exploring content. And that was precisely the hypothesis that I wanted to test with Pivot.
So, I set to the task of learning more about Pivot and, more importantly, designing my data structure and trade card design (a single item tile in Pivot is called a trade card) for MSDN Magazine articles. The result (which you can see from the magazine home page) looks as follows:
As I mentioned earlier, I ‘m going to approach this as a series – otherwise I’ll never get the time to write all this down. Here are the different posts that I’m planning – naturally, let me know if there’s something that you think I’ve forgotten and I’ll include it.
- Data and Design: The Chicken and Egg Problem
- Generating the Collection Data
- Creating the Trade Cards
- Hosting in Azure Blob Storage
- Instrumenting – What Have we Learned so Far?
I also created a few custom tools to help in cleaning up data and automating the collection creation process – I’ll share those either here or on the MSDN Code Gallery.