Graphs of Weekly Post-Count Growth
Welcome back to "Fun with Graphs"!
Today we'll use an earlier data set and make different graphs from it.
The data set is the same as in
post #61, namely weekly post counts. It's been a few weeks since that post was made, so there are a few additional weeks of data.
I used the same list of 36 Prolific Posters as before, and again broke them down into 3 groups of 12 members each.
The main change in these graphs is the filter applied to the basic data. Last time, "Weekly Total Posts" was too flat to show much, and "Posts per Week" was so jagged it was hard to read. For this graph, we show the cumulative growth in post counts from the start of the weekly samples to the most recent sample. In effect, each member gets their own "starting count", and the graph only shows growth from week to week. The first point on each graph is the member's weekly post count on the left-most date.
If a member makes a lot of posts for a week, then their line will slope up more steeply. If they make no posts for a week, then their line will be flat.
Since the members of each group have different post rates, you can't compare slopes between graphs. A steep line in one graph might not represent the same post rate in another graph. To get a sense of overall weekly growth, we need a single graph with all the Prolific Posters, which is shown last below.
Weekly Post Growth
View attachment 818497
View attachment 818498
View attachment 818499
Notice the Y axis numbers in Group 1 vs. Group 2. Due to the prodigious post rates of the top few members, the top of the Y range is several times larger in Group 1 than it is in Group 2.
The shape of each line in Group 1 is roughly the same, with a slope that varies little over time, suggesting that its members make roughly the same number of posts in a week. The orange line for
@I7guy and the light green line for
@Newtons Apple show a moderate change in slope near their ends, suggesting a change in their weekly post rates around that time.
The line shapes in Group 2 are more varied. In particular,
@cmaier and
@Dave245 both have significant upward slopes before the end of November, and are much flatter after that date. Group 3 shows an even more striking change, as the two members
@redheeler and
@D.T. both end with completely flat lines, indicating several weeks with no posts at all.
View attachment 818500
In this graph we
can compare slopes among all members, giving a better sense of how prolific
@Scepticalscribe is compared to everyone else. Even the steep lines from the graphs of Group 2 and Group 3 above disappear in a twisty maze of colored lines, all different.