COMP9321 25T1 Assignment 1 (15 marks)
Introduction
The NSW FuelCheck dataset is maintained by the NSW Government. It allows motorists
to access historical and live information about fuel prices across NSW. We have
downloaded the “FuelCheck Price History Jan 2025” file in February 2025 and are
storing it locally for your use as fuel.csv.
A dataset of Australian postcodes, with their latitudes and longitudes, is also available
from the user “Elkfox” on Github. This has been stored locally for your use in February
2025 as postcodes.json.
You're encouraged to review these and explore these datasets prior to attempting the
assignment. Please note that this data is publicly available, and we are not responsible
for political correctness, or other data inaccuracies, as this is purely a learning exercise
for data services engineering. Accompanying each question below you'll find:
Output: This includes the expected shape of the DataFrame to return.
Marking Considerations: These, along with the expected output, are provided to help
you stay on track. They serve as marking guidance to help you understand how we mark
the assessment.
Code Template
You must use the code template and rename the code template with your zID.
You must not modify the code template, except where indicated. For example, you
must not:
• Import additional third-party libraries.
• Modify the function signature for each question_X function.
• Modify the main function.
• Modify the log function.
• Disable or modify the calls to the log and plt.savefig functions within each
question_X function.
• Add any global variables.
You must setup a virtual environment as per the provided requirements.txt and
instructions here.
You must use either Python 3.11 (which is installed on CSE) or Python 3.13 (which is the
current release).
You must use pandas features to solve all question.
You must not iterate over the rows of any DataFrame. For clarity, any `for` or `while`
loop that runs proportionally to the number of rows in a DataFrame is iterating over the
rows of that DataFrame. This also holds for any Series, for example, iterating over one
column of a DataFrame. However, uou may iterate over DataFrame.columns, if you
feel it necessary.
You must not convert any DataFrame to a native data type (e.g. a list or dict) to process
the data.
You must not hard code any file paths within the code you write. These are specified in
the main function, which you must not modify, and are passed as parameters to the
question_X functions that you will complete. You must use these local variables
instead of hard-coded values.
You must not display any plots (e.g. using plt.show), as the code template is configured
to save your plot to disk.
You must not manually edit any datasets. During marking, a copy of the original CSV
and JSON files will be used.
You may import and use any of the Python 3.11 or 3.13 standard libraries. You may not
use any third-party libraries outside of what is provided in the virtual environment
setup.
You may write helper functions in the relevant code template section.


Part 1: Data Ingestion and Cleaning (4 marks)
Question 1: (1 marks)
Load the NSW FuelCheck dataset from the file fuel.csv into a Pandas DataFrame df1.
Handle the CSV input correctly, ensuring that erroneous data does not affect the
structure of the DataFrame.
Hint: Investigate why an extra column appears without any data in the header.
Return the DataFrame as df1.
Output: (60151, 8)
Marking considerations:
• [0.5 marks] The data in the DataFrame is correct.
• [0.5 marks] The shape of the DataFrame is correct, including the header.

Question 2: (1 marks)
Perform the following on the DataFrame from Question 1 df1, and return a new
DataFrame df2 with the following:
• Rename the column header “ServiceStationName” to “Name”.
• Ensure the text in the “Suburb” column is all upper case.
• Remove any rows for addresses not in NSW.
Output: (59256, 8)
Marking considerations:
• [0.5 marks] The above cleaning steps are performed correctly.
• [0.5 marks] The remaining data is not malformed in any way.

Question 3: (1 marks)
Load the postcodes dataset from the file postcodes.json into a new Pandas
DataFrame df3. Include all columns except for the “accuracy” column.
Return the DataFrame as df3.
Output: (16875, 6)
Marking considerations:
• [0.5 marks] The data in the DataFrame is correct.
• [0.5 marks] The header and removal of “accuracy” in the DataFrame is correct.

Question 4: (1 marks)
With the DataFrame df2 as the basis, create a column called “Latitude”, and another
column named “Longitude”. Populate these columns with the corresponding values as
per the postcodes DataFrame df3.
Use the “Postcode” and “PlaceName” fields in fuel.csv to match with postcodes.json.
If there are multiple matches, select the first option after sorting the matches
alphabetically. For example, 2018 is shared by Rosebery and Eastlakes, so Eastlakes
would be selected.
If there are no matches at all, leave the “Latitude” and “Longitude” fields blank.
Save the final DataFrame df4 as a CSV file df4.csv. Do not write the row names (index)
to the CSV file.
Return the final DataFrame as df4.
Output: (59256, 10)
Marking considerations:
• [0.5 marks] The data in the DataFrame is correct.
• [0.5 marks] The CSV file has the correct data and is named correctly.


Part 2: Data Exploration (2 marks)
Question 5: (2 marks)
Using the data in df4, create a new DataFrame df5, which will provide a summary of the
average fuel price per postcode and fuel type.
df5 should have a hierarchical index, where “Postcode” is the primary index and
“FuelType” is the secondary index.
The DataFrame should include the following:
• Index: Postcode
• Index: FuelType
• Column: AveragePrice
Return the final DataFrame as df5.
Output: (2584, 1)
Example:
Postcode FuelType AveragePrice
2088 E10 123.45
PDL 234.56
ULP91 345.67
…
2090 E10 456.78
PDL 0.00
ULP91 567.89

The postcodes should be sorted numerically, and the “FuelType” sorted alphabetically
within each postcode.
The final "AveragePrice" for each postcode and fuel type can be found by calculating:
• The average per-station and per-type daily fuel price, where there may be
multiple fuel readings per day for a particular fuel type and service station, which
are now aggregated to the service station level.
• The average per-postcode and per-type daily fuel price, where the daily averages
for each service station and fuel type are now aggregated to the postcode level.
• The final average per-postcode and per-type fuel price, where the daily averages
for each postcode and fuel type are now aggregated to a single value covering
the duration of the dataset.
If there are days without a calculated value for a particular fuel type and postcode,
exclude those from your final calculation.
In the final DataFrame df5, all “FuelType”s should be returned for each postcode, even
if that fuel is not sold in that postcode. The “AveragePrice” in that case would be 0.00.
Note that the “AveragePrice” should be calculated to 2 decimal places (2 d.p.). In this
case, 2 d.p. must always include all digits, even if they are .00.
Marking considerations:
• [0.25 marks] The columns have the correct headers
• [0.25 marks] The columns are sorted as requested.
• [0.5 marks] The indexes are correct.
• [1 mark] The “AveragePrice” calculation is correct and rounded to 2 d.p.

Part 3: Data Manipulation (2 marks)
Question 6: (1 marks)
Using the DataFrame df4 as the basis, create a column called “PriceChangeAverage”.
Populate the column with a 2 decimal place (2 d.p.) percentage difference of that row’s
fuel price with the corresponding average fuel price from df5. Note that in this case, 2
d.p. must always include all digits, even if they are .00.
Return the final DataFrame as df6.
Output: (59256, 11)
Example:
Assuming the following row in the previous question from df5:
2088, PDL, 234.56
Assuming your data in df4 contains the following row (some columns omitted for
example’s sake):
BP Mosman (9542),2088,PDL,207.9
Your df6 value for the “PriceChangeAverage” column would be -11.37.
Marking considerations:
• [0.25 marks] The column is added correctly.
• [0.25 marks] The sign (-/+) and decimal places are correct.
• [0.5 marks] The calculated “PriceChangeAverage” values are correct.

Question 7: (1 marks)
Using the DataFrame df6 as the basis, create a column called “PriceChangePrevious”.
Populate the column with the 2 d.p. difference of that row’s fuel price, with the previous
timestamp entry for that exact same fuel type and address. For the first row per unique
service station and fuel type in the DataFrame, set the field as zero. Note that 2 d.p.
must always include all digits, even if they are .00.
Return the final DataFrame as df7.
Output: (59256, 12)
Example:
Assuming your data in df6 contains the following row (some columns omitted for
example’s sake):
Astron Grafton,"105 BENT ST, SOUTH GRAFTON NSW 2460",SOUTH
GRAFTON,2460,ASTRON,P98,2025-01-01 09:09:09,184.9
Astron Grafton,"105 BENT ST, SOUTH GRAFTON NSW 2460",SOUTH
GRAFTON,2460,ASTRON,P98,2025-01-01 10:09:09,188.9
Your df7 value for the “PriceChangePrevious” column would be 4.00.
Marking considerations:
• [0.25 marks] The column is added correctly.
• [0.25 marks] The sign (-/+) and decimal places are correct.
• [0.5 marks] The calculated “PriceChangePrevious” values are correct.

Part 4: Data Visualisation (7 marks)
In this section, the most effective visualisations provide deep insights into the dataset.
You should also aim to achieve:
• Suitable choice of visualisation.
• Appropriate use of scale and colour.
• Appropriate inclusion of title, labels, legend, with suitable sizing.
• Visualisation is self-explanatory and informative.
• Image is correctly saved to disk, rather than shown.

Question 8: (3.5 marks)
Using the DataFrame df7, create a visualisation to answer the question: “Is it cheaper
to patron independent or franchised service stations?”
Save your visualisation as a PNG file (already setup in the code template) and explain
in a return variable answer8 what insights you have uncovered from this visualisation.
Note: you may use the “Brand” column to distinguish between Independent and all
other (Franchised) service stations.
Marking considerations:
• [1.5 marks] The visualisation provides deep insights into the research question.
• [1 mark] The visualisation is error free and of professional standard.
• [1 mark] The discussion has thorough and sound reasoning explaining how the
visualisation answers the research question.

Question 9: (3.5 marks)
Using the DataFrame df7, create a visualisation to answer the question: “Are
consumers in certain regions of NSW being unfairly charged for fuel compared to other
regions?”
Save your visualisation as a PNG file (already setup in the code template) and explain
in a return variable answer9 what insights you have uncovered from this visualisation.
Marking considerations:
• [1.5 marks] The visualisation provides deep insights into the research question.
• [1 mark] The visualisation is error free and of professional standard.
• [1 mark] The discussion has thorough and sound reasoning explaining how the
visualisation answers the research question.

学霸联盟