Description
Instructions
Four tasks will be listed here at different times during the semester. You should complete all tasks in a single jupyter notebook. This, along with relevant files like a README, should be in a single git repository synced with a hosting provider like GitHub [1]. That URL should then be submitted using the link on the Moodle page.
1. October 5th, 2020: Write a Python function called sqrt2 that calculates and prints to the screen the square root of 2 to 100 decimal places. Your code should not depend on any module from the standard library1 or otherwise. You should research the task first and include references and a description of your algorithm.
2. November 2nd, 2020: The Chi-squared test for independence is a statistical hypothesis test like a t-test. It is used to analyse whether two categorical variables are independent. The Wikipedia article gives the table below as an example [4], stating the Chi-squared value based on it is approximately 24.6. Use scipy.stats to verify this value and calculate the associated p value. You should include a short note with references justifying your analysis in a markdown cell.
White collar Blue collar No collar
Total
A B C D Total
90 60 104 95 349 30 50 51 20 151 30 40 45 35 150
150 150 200 150 650
1By the standard library, we mean the modules and packages that come as standard with Python. Anything built-in that can be used without an import statement can be used.
1
3. November 16th, 2020: The standard deviation of an array of numbers x is calculated using numpy as np.sqrt(np.sum((x – np.mean(x))**2)/len(x)) . However, Microsoft Excel has two different versions of the standard deviation calculation, STDEV.P and . The STDEV.P function performs the above calculation but in the calculation the division is by len(x)-1 rather than len(x) . Research these Excel functions, writing a note in a Markdown cell about the difference between them. Then use numpy to perform a simulation
demonstrating that the STDEV.S calculation is a better estimate for the standard deviation of a population when performed on a sample. Note that part of this task is to figure out the terminology in the previous sentence.
4. November 30th, 2020: NB – when I first posted this task, I accidentally wrote “k-means” where I meant to write “kNN” for k Nearest Neigh- bours. Because of this, I will allow either algorithm to be used and have extended the deadline by two weeks. Use scikit-learn to apply k-means clustering to Fisher’s famous Iris data set. You will easily obtain a copy of the data set online. Explain in a Markdown cell how your code works and how accurate it might be, and then explain how your model could be used to make predictions of species of iris.
STDEV.S
STDEV.S
2
References
[1] GitHub Inc., “GitHub,” https://github.com/.
Evidence of research performed on topic; sub- mission based on referenced literature, partic- ularly academic literature; evidence of under- standing of the documentation for any software or libraries used.
Environment can be set up as described; code works without tweaking and as described; code is efficient, clean, and clear; evidence of consid- eration of standards and conventions appropri- ate to code of this kind.
Evidence of planning and project management; pragmatic attitude to work as evidenced by well-considered commit history; commits are of a reasonable size; consideration of how commit history will be perceived by others.
Clear documentation of how to create an envi- ronment in which any code will run, how to pre- pare the code for running, how to run the code including setting any options or flags, and what to expect upon running the code. Concise de- scriptions of code in comments and README.
- [2] GMIT, “Quality Assurance Framework,” https://www.gmit.ie/general/quality-assurance-framework.
- [3] I. McLoughlin, “Using git for assessments,” https://github.com/ianmcloughlin/using-git-for-assessments/.
