Programming and Scripting

This project concerns the well-known Fishers Iris data set ^[3]. You must research the data set and write documentation and code (in Python ^[1]) to investigate it. An online search for information on the data set will convince you that many people have investigated it previously. You are expected to be able to break this project into several smaller tasks that are easier to solve, and to plug these together after they have been completed.

You might do that for this project as follows:

1. Research the data set online and write a summary about it in your README.
2. Download the data set and add it to your repository.
3. Write a program called analysis.py that:
   • outputs a summary of each variable to a single text file, saves a histogram of each variable to png files, and
   • outputs a scatter plot of each pair of variables.

It might help to suppose that your manager has asked you to investigate the data set, with a view to explaining it to your colleagues. Imagine that you are to give a presentation on the data set in a few weeks time, where you explain what investigating a data set entails and how Python can be used to do it. You have not been asked to create a deck of presentation slides, but rather to present your code and its output to them.

1

Minimum Viable Project

The minimum standard is a GitHub repository containing a README, a Python script, a generated summary text file, and images. The README should contain a summary of the data set and your investigations into it. It should also clearly document how to run the Python code and what that code does. Furthermore, it should list all references used in completing the project.

A better project will be well organised and contain detailed explanations. The analysis will be well conceived, and examples of interesting analyses that others have pursued based on the data set will be discussed. Note that the point of this project is to use Python. You may use any Python libraries that you wish, whether they have been discussed in class or not.

You should not be thinking of using spreadsheet software like Excel to do your calculations.

Advice for students

• Your git commit history should be extensive. A reasonable unit of work for a single commit is a small function, or a handful of comments, or a small change that fixes a bug. If you are well organised you will find it easier to determine the size of a reasonable commit, and it will show in your git history.
• Using information, code and data from outside sources is sometimes acceptable so long as it is licensed to permit this, you clearly reference the source, and the overall project is substantially your own work. Using a source that does not meet these three conditions could jeopardise your mark.
• You must be able to explain your project during and after its completion. Bear this in mind when you are writing your README. If you had trouble understanding something in the first place, you will likely have trouble explaining it a couple of weeks later. Write a short explanation of it in your README, so that you can jog your memory later.
• Everyone is susceptible to procrastination and disorganisation. You are expected to be aware of this and take reasonable measures to avoid them. The best way to do this is to draw up an initial straight-forward project plan and keep it updated. You can show the examiner that you have done this in several ways. The easiest is to summarise the project plan in your README. Another way is to use a to-do list like GitHub Issues.
• Students have problems with projects from time to time. Some of these are unavoidable, such as external factors relating to family issues or illness. In such cases allowances can sometimes be made. Other problems are preventable, such as missing the submission deadline because you are having internet connectivity issues five minutes before it. Students should be able to show that up until an issue arose, they had completed a reasonable and proportionate amount of work and took reasonable steps to avoid preventable issues.
• Go easy on yourself this is one project in one module. It will not define you or your life. A higher overall course mark should not be determined by a single project, but rather your performance in all your work in all your modules. Here, you are just trying to demonstrate to yourself, to the examiners, and to prospective future employers, that you can take a reasonably straight-forward problem and solve it within a few weeks.

References

• Python Software Foundation. Welcome to python.org. https://www.python.org/.
• Quality assurance framework. https://www.gmit.ie/general/qualityassuranceframework.
• UC Irvine Machine Learning Repository. Iris data set. http://archive.ics.uci.edu/ml/datasets/Iris.