As I was writing my community engagement post, I allowed myself to get pulled into a rabbit hole. That rabbit hole was data analysis. My preferred Professional Learning Network is the National Science Teachers Association listserve, as detailed in my previous post where I reflected on my experiences with various forms of PLN’s. Analyzing user data is a rich field of study and is becoming increasingly important as Big Data continues to amass amazing sums of user information, learning to handle and manipulate it, and draw justifiable conclusions.
An analysis of my NSTA activity
Someone is probably going to chime in and tell me that there was an automated way to do this, and normally I’d let that bother me, but I’m learning to see the value in doing things manually sometimes. There is a lot of good thought that happens during repetitious tasks, and it was actually a very fun (and sometimes cringe-worthy) trip down memory lane.
I searched for emails from the NSTA Chemistry listserve (firstname.lastname@example.org) that contained my last name, which is unique in the listserve, and set the date range within 1 year of today. I repeated this for each listserve, as well as all listserves combined.
For the nerds out there, the full search term (is this Boolean?) was:
(to:(email@example.com OR firstname.lastname@example.org OR email@example.com), lenz) after:2015/4/14 before:2016/4/15
After about 20 minutes of clicking around, I managed to generate a nice little table. I also quickly popped out a nice graph, which was a nice reminder that Google Sheets continues to improve (I remember trying to make graphs a year or two ago on Sheets and it was terrible!).
This brought up a few interesting questions, as good research often does.
1.) What was responsible for the downturn in my NSTA involvement this year?
2.) Was Twitter siphoning off some of the involvement I normally would have given to NSTA?
3.) Is involvement in PLN’s a zero sum game?
In order to look at the relationship between NSTA and Twitter, I needed to get data. A quick search brought up Twitter Analytics, which is a gold-mine of data, though if you want to play with it yourself, you’ll need to be patient. The default analysis tools are neat, but they only analyze month-by-month, and really just compare the current month to the one previous.
To get a longer-view, I needed to get the data myself.
The good news is: You can download the raw data directly as a .CSV file! The bad news: You can only download it in ~80 day windows. So, click, click, click. After about 5 minutes, I had a big ugly spreadsheet that reached back 3 years, to the beginning of my Twitter usage.
Obviously there is a lot of data here. I only copied about 1/4 of the columns, and really all I am interested in is the timestamp column. What I’d like to do is make a histogram of the frequency of posts per day. Simple, right? Well, after messing around with it for about 30 minutes and coming up with all kinds of useless and not what I wanted! graphs, I’m calling it quits for now.
To answer some of my own questions, I believe that a.) involvement in PLN’s is not a purely zero sum game, but there are only 24 hours in a day, and realistically speaking, you can only spend X numbers of them collaborating and sharing with others. In reality, that value of X is frustratingly small….I’d say it should probably phrased more accurately as 0.0ox!
It is interesting to note that my involvement in the biology listserve dropped off very quickly in at the end of 2009, and it took me a while to remember that this was when I stopped teaching biology, and allowed chemistry to become my sole focus. The uptick in physics reflects my increasing interest (and teaching load) in this subject. Overall, I am involved in about 1 thread per week.
What this data doesn’t show is the amount of lurking I do, which is enormous! I glean worksheets, lab ideas, photos, jokes–even job leads. This is a wonderful resource and I was happy to see its importance shown in graphical form.