I love Boston. I was there for a conference, went to MIT and used their toilet.
It’s a world-class school, but it turns out that Massachusetts in general has a proven reputation for success amongst students – consistently:
Why is this and what can be learnt from their example ?
I’ll present several visualisations below (click for full view), taken from the wealth of data appreciatively and publicly obtained from http://profiles.doe.mass.edu/ , which attempts to boil down the factors of success.
Figure 1: School locations, enrollment size, and student success level
This map shows the locations of the high-schools within the state, their distribution and relative size in terms of enrollment, and lastly as an in indication of each schools relative success in having their students going onto to graduate (finish college and go on to University), as indicated by the colour scale: red 0-50%, yellow 50-75%, green 75-100%.
Figure 2: Factors influencing Students proceeding to Advanced Placement
The AP is the Advanced Placement programme used to identify and enable students demonstrating gifts that, given the correct guidance, may well see them do well in an advanced placement at University.
The dashed trend line above shows the correlation between achievement of an AP score of 3 or more (being the nominal “pass” rate), and graduation rate. The dots are then coloured according to the percent of students who went on to college following graduation. The line shows a positive correlation between the two variables, though with an R-square value of 0.52, meaning that while the variables maintain covariance, the correlation is predictive only approximately half the time.
Figure 3: Student-Teacher ratio and influence on Graduation Rate
Average Student-Teacher Ratio vs. average Graduation Rate. The data is filtered on average of % to college and average of % AP3+, and is shown as a heat density, which shows clustering by amount of overlay amongst the plot points.
Figure 4: Effect of expenditure on graduation rate
This chart to me was probably the most interesting one. And not just because it allowed me to my use my “Aliens 2 movement detector” temperature-divergent heat mapping cluster graphing software again, but because it immediately reveals patterns, outliers, and clusters that would never be apparent gazing over numbers – the beauty and purpose of exploratory data analysis and visualisation. The commentary on the annotations tell the story that became apparent to me upon seeing the results.
