salmanjavaid • over 6 years ago
How do students perform across Arts and Science subjects and their interdependencies?
Sometime back, I got a dataset with over 50,000 transcripts for 9th and 10th class examinations from Lahore Board of exams. I had it parsed, and the data is with me now in excellent condition detailing each students' marks in each of the subjects.
Here is what I have in mind, and will love some feedback. But before, let me narrate an analogy to get you onboard:
So you are a banker, a banker supposed to grant loans to individuals. And I come to you one day and apply for a loan. Now you call me back, and ask me to submit details regarding my job, my work experience, my education, my age bracket, my dependents, my belongings, and my previous credit score.
Now you need to decide if you are going to provide me loan. How do you do that? Well, you sit down, and do some calculations on the details I provide to you.
I hope you are with me till here. Now here is where it gets interesting. There are algorithms already out there that find relationships in data, how different variables relate to each other, what are the probabilities between them, for example, in the preceding example, consider the huge wealth of data from previous applications that you have accrued over time, and you run the algorithm on it and try to find relationships between various variables.
The same way you can run the algorithm on this data and try to find how the students who perform well in mathematics perform in say English or Urdu, or Physics, or Computer Science. You can even create multiple dependencies, e.g., between English, Urdu, Physics, and Computer Science, and try to find the underlying relationships between data.
Even more, you can even introduce further variables in this data if you have information regarding teachers, the teacher's performances, location of schools. The possibilities endless.
Please ping me if I have confused you, and I apologize for that. And if you are interested, lets get started.
Comments are closed.