Today I was attending a data visualization workshop organized by the KNAER Data Visualization project in collaboration with AERO. I was one of the presenters, including Paul Anisef, Rob Brown, Chris Conley and Cosmin Marmureanu. It was the first time I met with them in person although we have been working together for a few months already on data visualization.
The workshop went quite well, with around 60 participants mostly from school boards and the ministry of education. Rob started by introducing the history of this data visualization project. After that, Chris took over and gave a nice introduction to a number of important principles of data visualization. Then he guided everybody working through a few neat data visualization techinques in Excel, including pyramid chart, slope chart, thermometer chart, and interactive charts. As a person who mainly used Excel for bar charts when coming to data visualization, it was amazing to learn those new techniques.
After the Excel session was done, Cosmin who has been working extensively on GIS visualization of educational data introduced us to a few mapping tools including Google Fusion Tables, ArcGIS and Quantum GIS. Putting data on top of maps could become really powerful as I saw at the Toronto Open Data Day event a few weeks ago. Mashing up educational data on top of maps could provide fresh ways of looking at the data that cannot be found elsewhere, and could help communicate information strongly to impact planning and policy making. However, interpreting information from a map is always tricky and needs special caution, because there is always rich information behind or attached to the map and either the presenters or the audience could easily become biased.
After lunch, I led a session focusing on data visualization with R. It was the first time I was giving a presentation solely focusing on R since I learned it last year in a course on Coursera. It was also the first time to present to an audience mainly from school boards. So the experience was quite interesting. As most of the participants were not familiar with R, I tried to make everything simple and straightforward, building visualizations step by step by adding one new element each time. The presentation mainly focused on ggplot2, and briefly mentioned Shiny and knitr at the end. By showing what you can do with only one line of code in R, I was trying to convey how powerful and flexible R could become especially in exploring data. The demo I did mainly focused on some ordinary data visualization techniques such as barplot, boxplot, histogram, and scatterplot. I also included a few slides about data manipulation and transformation in R as well as more advanced stuff like heatmaps and geographic maps. After this demo was done, Chris showed how to make venn diagram and Google motion chart in R, which was very cool. I’ve posted my slides and handout online. (They themselves were written with R in this file and compiled into slides with Slidify and to handout with knitr.)
We left the day after Paul wrapped up the workshop discussing future possibilities of this project and knowledge mobilization work to be done.