For a d3 bar plot visit http://www.jaredlander.com/plots/PizzaPollPlot.html.
require(rjson) require(plyr) pizzaJson <- fromJSON(file = "http://jaredlander.com/data/PizzaPollData.php") pizza <- ldply(pizzaJson, as.data.frame) head(pizza)
## polla_qid Answer Votes pollq_id Question ## 1 2 Excellent 0 2 How was Pizza Mercato? ## 2 2 Good 6 2 How was Pizza Mercato? ## 3 2 Average 4 2 How was Pizza Mercato? ## 4 2 Poor 1 2 How was Pizza Mercato? ## 5 2 Never Again 2 2 How was Pizza Mercato? ## 6 3 Excellent 1 3 How was Maffei's Pizza? ## Place Time TotalVotes Percent ## 1 Pizza Mercato 1.344e+09 13 0.0000 ## 2 Pizza Mercato 1.344e+09 13 0.4615 ## 3 Pizza Mercato 1.344e+09 13 0.3077 ## 4 Pizza Mercato 1.344e+09 13 0.0769 ## 5 Pizza Mercato 1.344e+09 13 0.1538 ## 6 Maffei's Pizza 1.348e+09 7 0.1429
require(ggplot2) ggplot(pizza, aes(x = Place, y = Percent, group = Answer, color = Answer)) + geom_line() + theme(axis.text.x = element_text(angle = 46, hjust = 1), legend.position = "bottom") + labs(x = "Pizza Place", title = "Pizza Poll Results")
But given this is live data that will change as more polls are added I thought it best to use a plot that automatically updates and is interactive. So this gave me my first chance to need rCharts by Ramnath Vaidyanathan as seen at October’s meetup.
require(rCharts) pizzaPlot <- nPlot(Percent ~ Place, data = pizza, type = "multiBarChart", group = "Answer") pizzaPlot$xAxis(axisLabel = "Pizza Place", rotateLabels = -45) pizzaPlot$yAxis(axisLabel = "Percent") pizzaPlot$chart(reduceXTicks = FALSE) pizzaPlot$print("chart1", include_assets = TRUE)
Unfortunately I cannot figure out how to insert this in WordPress so please see the chart at http://www.jaredlander.com/plots/PizzaPollPlot.html. Or see the badly sized one below.
There are still a lot of things I am learning, including how to use a categorical x-axis natively on linecharts and inserting chart titles. I found a workaround for the categorical x-axis by using
tickFormat but that is not pretty. I also would like to find a way to quickly switch between a line chart and a bar chart. Fitting more labels onto the x-axis or perhaps adding a scroll bar would be nice too.
Jared Lander is the Chief Data Scientist of Lander Analytics a New York data science firm, Adjunct Professor at Columbia University, Organizer of the New York Open Statistical Programming meetup and the New York and Washington DC R Conferences and author of R for Everyone.