While the Hubway Data Challenge has been an opportunity to explore some of the newer capabilities in MicroStrategy 9.3, it was also an opportunity to work with a new data set and a get some practice modeling in geolocation data. Working with other developers, and seeing some of the submissions has informed my understanding of what tools are out there and what other people are doing with commercial and open source visualization tools. The concept of animating this data on top of a map has particular appeal is appropriate for this type of data.
This submission caught my attention, but there are a few others that have been submitted in recent days.
For me I was committed to sticking with the 10.31 deadline even though the deadline was extended. I had no shortage of ideas, but ultimately I had to pick something to show and share. The new network diagrams in 9.3 are clean looking and visually pleasing. So, I picked one bike (the mayor’s) and graphed it’s usage. I set the export option for dashboards to be .pdf (as opposed to .mht) and was able to create a portable document from the Visual Insight interface.
The .pdf file is available here.
The other thing I wanted to try with the data was something using the maps. It took some trial and error and an angry email to Google before I realized that I didn’t in fact have to pay $10,000 for a Google maps API key. To get the data in a format that would work for the maps, I decided to create a running sum of the trips by station over time. There were two things I had to do within MicroStrategy to make this work:
1. Set the VLDB property to retain all of the lookup table information
2. Create a running sum metric to accumulate trip information:
Within the grid report the output had to look something like this before I could build a cube off of it:
The next step was to create a Visual Insight document using the map style. I had tried to use the density version but my data was not disparate enough to use close up, so I used the traditional map and used bubbles to illustrate the data. I set the geographic attribute to be the start station, and because I modeled in the lat and long as part of the attribute form (as latitude and longitude data types) Visual Insight recognized the data. The last step was to put the start date into the page by and from there I could take advantage of the play button.
I used the Quicktime player on my laptop to record my screen and then I used iMovie to speed up the movie to condense something that took ~6 minutes in MicroStrategy to play in about 24 seconds. This became the basis for my second submission to the Hubway Data Challenge. I posted this on Youtube:
My overall impression with 9.3 has been favorable so far. The cleaned up UI is pleasing, the interface is faster, and the Visual Insight enhancements not only look good, but they work pretty well too. Having spent some time with other developers at the hack day event I realized just how many people are using R, and with the integration of R into MicroStrategy there are certainly some possibilities to consider there. I know from past experience that the number of freely available R packages is improving, and that the corresponding library of visualization for R is growing. This offers many possibilities for MicroStrategy development and deployment, and certainly extends the platform in ways that I am looking forward to exploring.