Today is the one year anniversary of the NYC R Statistical Meetup. Starting as a small group meeting in a crowded conference room, Josh Reich established the meetup as the premiere gathering of data geeks in the tri-state area. Over the past year we have had 10 meetups, and two more upcoming on the calendar, covering a wonderfully varied set of topics on the application of R.
All of this activity has also generated a large set of data, and what better way to celebrate anniversary of our group than to analyze that data using R?
Meetup Descriptions Word Cloud
First, I downloaded all of the meetup descriptions into a text file (yes, I copied and pasted them) and created a word cloud.
This was actually very easy in R, taking only about 40 lines of code.
There is probably a better way to specify the X and Y coordinates in order to minimize word overlap, but after some trial and error
runif gives decent results. The above cloud actually provides interesting insight into our group. As you can see, we are focused on analysis using specific R packages, particularly with respect to graphics. Also, the Conways remain a dominate force in all things data and R.
Member Activity Time-Series
Meetup.com provides organizers with lots of data related to a group’s activity. They provide information on the growth of the group, as well as the number of people joining and RSVPing events. Below is a combined chart of this data.
The data was generated by Meetup.com, so here the trick was getting everything into a single data frame and dealing with missing data and formats.
UPDATE: I tweaked the chart a bit to show the data more clearly, as well as have the true number of members along with the total number of “active” members. For some reason the data provided on group size does not start until mid-August 2009, so I have used the number of people joining the group before then as an estimate.
Note that the total number of members has been divided by 10 for scaling. I am not completely satisfied with this graphic, but it does provide a good illustration of the cyclical nature of member activity and the constant rate of growth for our group. Also, it is nice to see the number of RSVPs increasing—perhaps we are writing better descriptions.
The code and data used to generate these charts can be downloaded in the ZIA Code Repository. As always, I welcome other analyses and suggestions.