CSCA08H Assignment 3: Hypertension and Low Income in Toronto Neighbourhoods

CSCA08H Assignment 3: Hypertension and Low Income in Toronto Neighbourhoods

Goals of this Assignment

QQ1703105484

In this assignment, you will practise working with files, building and using dictionaries, designing functions using the Function Design Recipe, reading documentation, and writing unit tests.

Background

A commonly-held belief is that an individual's health is largely influenced by the choices they make. However, there is lots of evidence that health is affected by systemic factors.

Health researchers often study the relationships between an individual's health outcomes and factors related to their physical environment, social and economic situations, and geographic location. Studies such as this one investigate how a particular health outcome (living with hypertension) are tied to a systemic factor (the income level of a country).

In this assignment, you will write code to assist with analysing data on the relationship between hypertension (also known as high blood pressure) and income levels in Toronto neighbourhoods. The data you will work with is real data, however we have simplified it somewhat to make this assignment clearer for you.

A note on math and stats

The data analysis that your code will do will include some statistical analysis that we have not talked about in the course. You do NOT need to understand the underlying statistics to complete this assignment. The code you write will do some simple mathematical operations, like adding up some numbers, or finding ratios using division. We will use Pearson correlation for the more advanced analysis and you will use existing functions that we have imported for you.

You will need to take a look at the examples of these functions in order to figure out what arguments you need to pass to them, and what types of data they return, but you do not need to understand how they work in any detail.

Correlation is a single coefficient expressing the tendency of one set of data to grow linearly, in the same or opposite direction, with another set of data. This is done by comparing whether points that have been paired between the two sets are similarly greater or less than than their set's respective averages.

For example, if we wanted to compare whether for students in the class, age is correlated with height, we would have two sets of data, birth date (which we could express as, say, number of weeks old for finer granularity), and heights.

Numbers from each set are ordered in the same way so that each height value corresponds to the age value for the same student. What is nice about the correlation metric we are using, is that it is normalised to be between -1 and 1, with these values giving us a nice human interpretation. A value of 1 means that the points make a straight line. In our example, this means, for some increase in age, we have a consistent increase in height. Similarly, a value of -1 is the same relationship but with a flip of direction, where older students would be shorter than younger ones. Finally, a value of 0 would say that there is no consistent increase or decrease in height for a change in age. We will use this to investigate the relationship between low income rates and hypertension, for any tendency to increase or decrease together.

If you are a statistics person, keep in mind that the learning goals of the assignment are about writing code using what we've learned in the course, not about doing a proper statistical analysis.

Dataset descriptions

This assignment uses data files related to one of the two variables of interest (i.e., hypertension data or income data). The files are CSV (comma separated values) files, where each column in a line is separated by a comma. You can assume there are no commas anywhere else in the files, other than to separate columns, and that any file given is in the correct format. The two file types are described below.

Neighbourhood hypertension data files

The first row in a neighbourhood hypertension file contain header information, and the remaining rows each contain data relating to hypertension prevalence in a particular Toronto neighbourhood.

Here is a description of the different columns of the dataset. Notice the use of constants and carefully study the starter file constants.py.

Neighbourhood income data files

The first row in a neighbourhood income data file contains header information, and the remaining rows each contain data about low income status.

Here is a description of the different columns of the dataset. Notice the use of constants and carefully study the starter file constants.py.

Neighbourhood names and ids are the same between our hypertension data files and our low income data files. However, the total population of a neighbourhood can be different between the two data files, as they were collected at different times.

The CityData Type

The code you will write for this assignment will build and then use a dictionary that contains hypertension and low income data about neighbourhoods in a city. This section describes the format of that dictionary.

Key/value pairs in a CityData dictionary

Each key in a CityData dictionary is a string representing the name of a neighbourhood. As is necessary for dictionary keys, all neighbourhood names will be unique.

The values in a CityData dictionary are dictionaries containing information about a neighbourhood. These neighbourhood data dictionaries contain specific keys that label a neighbourhood's data.

Format of the inner dictionaries

A dictionary that is a value in a dictionary of type CityData has the following key/value pairs. Notice the use of constants and carefully study the starter file constants.py.

An example CityData dictionary

The following is an example of a CityData dictionary. We have also provided this dictionary for you to use in your docstring examples and other testing in the starter code file. Note that we have formatted the dictionary below for easier reading, however you will not see this formatting in your code.

{'West Humber-Clairville': {
    'id': 1,
    'hypertension': [703, 13291, 3741, 9663, 3959, 5176],
    'total': 33230,
    'low_income': 5950},
'Mount Olive-Silverstone-Jamestown': {
    'id': 2,
    'hypertension': [789, 12906, 3578, 8815, 2927, 3902],
    'total': 32940,
    'low_income': 9690},
'Thistletown-Beaumond Heights': {
    'id': 3,
    'hypertension': [220, 3631, 1047, 2829, 1349, 1767],
    'total': 10365,
    'low_income': 2005},
'Rexdale-Kipling': {
    'id': 4,
    'hypertension': [201, 3669, 1134, 3229, 1393, 1854],
    'total': 10540,
    'low_income': 2140},
'Elms-Old Rexdale': {
    'id': 5,
    'hypertension': [176, 3353, 1040, 2842, 948, 1322],
    'total': 9460,
  'low_income': 2315}}

The sample CityData dictionary above represents hypertension and low income data for five neighbourhoods: West Humber-Clairville, Mount Olive-Silverstone-Jamestown, Thistletown-Beaumond Heights, Rexdale-Kipling, and Elms-Old Rexdale.

Let's take a closer look at the data for Elms-Old Rexdale. This neighbourhood is represented by the key/value pair where the key is 'Elms-Old Rexdale'. The id of this neighbourhood is 5. The hypertension data for this neighbourhood is as follows: 3353 people are between the ages of 20 and 44, 176 of whom have hypertension. There are 2842 people between the ages of 45 and 64, 1040 of whom have hypertension, and there are 1322 people aged 65 and up, 948 of whom have hypertension. The low income data for this neighbourhood is that 2315 people are classified as low income, from a total population of 9460 people.

Note that the totals do not match between the low income and the hypertension data — this is because the low income data was collected before the hypertension data, and the size of the neighbourhoods changed. For the purposes of this assignment, we will assume the collection of these two datasets is close enough in time to compare them to each other. You do not need to do anything about these differing totals, other than to make sure you are using the correct total when computing rates, as described later.

Age standardisation

This section describes the process of age standardisation that we will use in this assignment to perform a more accurate analysis. Note that we have given you a function that computes the age standardised rate from the raw rate (described in Task 3). This section is for your information only; we have already implemented this for you.

Our dataset will let us calculate the rate of hypertension in each Toronto neighbourhood. One complicating factor is that different neighbourhoods have different age demographics. For example, the Henry Farm neighbourhood has a significantly lower proportion of 65+ residents than Hillcrest Village. And because people aged 65+ have a higher overall rate of hypertension, this demographic difference alone would cause us to expect to see a difference in the overall hypertension between these neighbourhoods.

So because we care about the impact of low income status on hypertension rates, we want to remove the impact of different age demographics between the neighbourhoods. To do so, we will use a process called age standardisation to calculate an adjusted hypertension rate that ignores differences in ages. This process involves the following steps for each neighbourhood:

1. First, we'll calculate the hypertension rate within each of the following age groups: 20-44, 45-64, and 65+. We'll report these rates as percentages, which you can think of as being the number of cases of hypertension per 100 people aged 20-44 / 45-64 / 65 and up.
2. Then, we'll pick one standard population with certain numbers of people in these age groups. For the purpose of this assignment, we'll use the total Canadian population from the 1991 census:
3. Then, we'll use the neighbourhood rates to calculate the hypothetical number of people in the standard population who would have hypertension. For example, if the rates for neighbourhood X were 20% of 20-44, 30% of 45-64, and 66% of 65+, the total number of people with hypertension in the standard population would be 2,239,966 + 1,609,760 + 2,092,180 = 5,941,906.
4. Finally, divide this number of people with hypertension by the total size of the standard population, yielding a final percentage 5,941,906 / 19,735,665 x 100 or approximately 30%. This percentage is the age standardised rate for the neighbourhood.

If you are interested, you can read more about age standardised rates here.

Required Functions

In the starter code file a3.py, follow the Function Design Recipe to complete the functions described below.

You will need helper functions (i.e., functions you define yourself to be called in other functions) for some of the required functions, but likely not for all of them. Helper functions also require complete docstrings with doctests. We strongly recommend you also follow any suggestions about helper functions in the table below; we give you these hints to make your programming task easier.

Some indicators that you should consider writing a new helper function, or using something you've already written as a helper are:

• Rewriting code to solve a task you have already solved in another function
• Getting a warning from the checker that your function is too long
• Getting a warning from the checker that your function has too many nested blocks or too many branches
• Realising that your function can be broken down into smaller sub-problems (with a helper function for each)

For each of the functions below, other than the file reading functions in Task 1, write at least two examples in the docstring. You can use the provided SAMPLE_DATA dictionary, and you should also create another small CityData dictionary for examples and testing. If your helper function takes an open file as an argument, you do NOT need to write any examples in that function's docstring. Otherwise, for any helper functions you add, write at least two examples in the docstring.

Your functions should not mutate their arguments, unless the description says that is what they do.

Assumptions

Assume the following about the data:

• All neighbourhood ids and names are unique, and will appear the same in all data files. That is, no neighbourhood will have a different id between files, or a different name.
• In all tasks except Task 1, the dictionary argument will have both hypertension and low income data for every neighbourhood. That is, it will be a valid CityData dictionary.
• All float values should be left as is; do not round any of them.

Using Constants

The starter code contains constants in the file constants.py that you should use in your solution for the list indices and key identifiers for the CityData dictionary as well as the column numbers for the input files. You may add other constants if you wish, but DO NOT place them in the file constants.py: instead put them in the a3.py file.

Task 1: Building the data dictionary

In this task, you will write functions that read in files and build the dictionary of neighbourhood data. You will write two functions — one that adds hypertension data to a dictionary, and one that adds low income data. You will almost certainly also need to define one or more helper functions to help you solve this task.

These functions will be used to build a CityData dictionary, however the dictionary that is passed to the functions may not yet contain all of the data.

To illustrate this, we have provided two small data files. After passing the same dictionary to both functions with each of those small files, the dictionary should be a CityData dictionary that contains the same information as the provided SAMPLE_DATA dictionary. Using the small hypertension file and an empty dictionary as arguments to get_hypertension_data, the result should be that the dictionary now contains the hypertension data as in SAMPLE_DATA, but not the low income data.

 {'West Humber-Clairville':
  {'id': 1, 'hypertension': [703, 13291, 3741, 9663, 3959, 5176]},
  'Mount Olive-Silverstone-Jamestown':
  {'id': 2, 'hypertension': [789, 12906, 3578, 8815, 2927, 3902]},
  'Thistletown-Beaumond Heights':
  {'id': 3, 'hypertension': [220, 3631, 1047, 2829, 1349, 1767]},
  'Rexdale-Kipling':
  {'id': 4, 'hypertension': [201, 3669, 1134, 3229, 1393, 1854]},
  'Elms-Old Rexdale':
  {'id': 5, 'hypertension': [176, 3353, 1040, 2842, 948, 1322]}}

Similarly, using the small low income file and an empty dictionary as arguments to get_low_income_data, the result should be that the dictionary now contains the low income data as in SAMPLE_DATA, but not the hypertension data.

 {'West Humber-Clairville':
  {'id': 1, 'total': 33230, 'low_income': 5950},
  'Mount Olive-Silverstone-Jamestown':
  {'id': 2, 'total': 32940, 'low_income': 9690},
  'Thistletown-Beaumond Heights':
  {'id': 3, 'total': 10365, 'low_income': 2005},
  'Rexdale-Kipling':
  {'id': 4, 'total': 10540, 'low_income': 2140},
  'Elms-Old Rexdale':
  {'id': 5, 'total': 9460, 'low_income': 2315}}

A complete CityData dictionary will have been passed to both functions. See the sample usage at the end of the starter code file for an example of how both functions are used to build a CityData dictionary.

Note: While this is the first task, it is not necessarily the easiest. If you are stuck while working on this task, we suggest moving on to other tasks and coming back to this later.

Recall that TextIO as the parameter type means the file is already open.

Task 2: Neighbourhood-level Analysis

Task 3: Finding the Correlation

Task 4: Order by Ratio

Task 5: Required Testing (unittest)

Write and submit a unittest file for the get_bigger_neighbourhood function. We have provided starter code in the test_a3.py file. We have included one test that you can use as a template to write your other test methods. For each test method, include a brief docstring description specifying what is being tested. Do not write examples in the docstrings. Your set of tests should all pass on correct code, and your tests should be thorough enough that at least one of them will fail on a buggy version of the function. There is no required number of tests; we will mark your tests by running them on the correct code as well as several buggy versions.

Files to Download

Download a3.zip which contains starter code (a3.py and test_a3.py), the checker (a3_checker.py together with the helper file checker.py and folder pyta), and two sizes of each type of data file.

Marking

These are the aspects of your work that will be marked for Assignment 3:

• Correctness (70%): Your functions should perform as specified. Correctness, as measured by our tests, will count for the largest single portion of your marks. Once your assignment is submitted, we will run additional tests, not provided in the checker. Passing the checker does not mean that your code will earn full marks for correctness.
• Testing (15%): Your test suite will be checked by running it on incorrect/broken implementations. Your tests should all pass on a correct version of the function, and at least one should fail on each of our broken implementations.
• Coding style (15%):
  • Make sure that you follow Python style guidelines that we have introduced and the Python coding conventions that we have been using throughout the semester. Although we don't provide an exhaustive list of style rules, the checker tests for style are complete, so if your code passes the checker, then it will earn full marks for coding style with one exception: docstrings may be evaluated separately. For each occurrence of a PyTA error, one mark (out of 20) deduction will be applied. For example, if a C0301 (line-too-long) error occurs 3 times, then 3 marks will be deducted.
  • If you encounter PyTA error R0915 (too-many-statements), that indicates that your function is too long (more than 20 statements long). In that case, introduce helper functions to do some of the work — even if the helpers will only be called once. Your program should be broken down into functions, both to avoid repetitive code and to make the program easier to read.
  • All functions, including helper functions, should have complete docstrings including preconditions when you think they are necessary.
  • Also, your variable names and names of your helper functions should be meaningful. Your code should be as simple and clear as possible.

What to Hand In

The very last thing you do before submitting should be to run the checker program one last time.

Otherwise, you could make a small error in your final changes before submitting that causes your code to receive zero for correctness.

Submit a3.py and test_a3.py on MarkUs by following the instructions on the course website. Remember that spelling of filenames, including case, counts: your file must be named exactly as above.