Making Big Data more accessible by getting rid of the numbers

Schools are data generating machines. Data drop deadlines are red-letter days in the school calendar and students are systematically measured on test results, effort, homework submission, behaviour, punctuality, predicted grades, etc. We are in the midst of the education big data movement. Some have issue with what they consider obsessively measuring in the hopes that it fattens the minds of our students. I see a different problem: how we present this data to teachers so that it is meaningful and useful, instead of a deluge of numbers and codes.

My research question (in its latest iteration) is “can changing the way we present data to teachers positively impact how useful they find it?” I began by informally interviewing colleagues from a few different departments and career stages about their attitude to the use of data in their planning. All of them spoke about how important data is for informing their teaching, but there were two common negative themes: it can be difficult to organise the data in a meaningful way, and they could become intimidated by the sheer volume of numbers. Following these interviews, I researched different ways of displaying data, including the fascinating world of infographics. I settled on the use of “bubble charts” for a number of reasons, including that Microsoft Excel can be used to generate them and I wouldn’t need to purchase any additional software.

Bubble charts allow you to present data that includes four different variables. There is the typical x and y-axes for two variables, a z-axis (the area or diameter of the bubble) for a third variable, and the bubble can be coloured for a fourth variable. I developed a number of different formats and am grateful to the verbal and emailed feedback from my colleagues. I finally settled on the example below, where the x-axis lines are individual students, the y-axis lines are individual subjects, the size of the bubble is how many behaviour points the student has racked up in that subject and the colour represents their progress vs target for that subject on their latest report. Importantly, even though the whole cohort graph presents almost 2000 data points, it does not include a single number. I added a number of “buttons” which make the data easily sortable and then sent copies to my colleagues.

The y-axis shows the data for (from bottom to top) English, maths, science. The student names have been replaced with numbers in this example.

The response so far has been very positive and the biggest surprise has been how confidently the sampled teachers and department heads have begun interrogating the data using these charts (is student X in the wrong set? What is happening in subject Y for student Z? Where is best practice for progress by disadvantaged group X?). Another surprise has been how colleagues assume that it was a huge amount of work to put the charts together, when in reality I have not collected or created any new student data – I have simply taken the existing data and presented it in a new, and easily reproducible way using existing software tools.

The next steps in this project is are twofold. Firstly, I will generate charts for tutors to use in their upcoming academic mentoring sessions with their year 10 students. These charts will be individualised for each student and reflect their effort, behaviour and progress in all their subjects (again all existing data). This will be followed by a simple survey for the students and their teacher-mentors about the utility and usability of these charts vs the numerical tables typically used. Secondly, I will introduce heads of department to the charts for use, if desired, in their setting plans for next year. Again, a survey will be used to gauge utility and usability with an eye to another iteration of improvement and refinement.

Sam Craven

Ely College