Data Visualisation Foundation Micro-credential – Project

1.Executive Summary
The global impact of the COVID-19 pandemic has been significant, affecting lives and disrupting daily routines.
Australia has also faced its share of challenges, experiencing multiple waves of the virus. The main purpose of this
report is to showcase the advanced capabilities of Power BI as a data visualization tool. To demonstrate this, the
report extensively analyses the COVID-19 situation in Australia and its states, utilizing the Power BI report built on
two data sets: 'Covid-19' and 'States'.

This report offers a comprehensive overview of the Power BI report itself. It delves into the characteristics of the
datasets, the data cleaning and formatting processes employed, and the establishment of relationships between the
different data tables. It explores the selection of visuals on the report page, providing insight into the reasoning
behind their choices. The report also explains the adjustments made to the visual formatting to enhance the overall
appearance and interactivity, along with the utilization of various filters.

Furthermore, the report presents a valuable summary of the key findings derived from the COVID-19 analysis. It
highlights trends, patterns, and essential insights gained from the examination, showcasing the practical applications
of data visualization and analysis in real-world contexts.

2. Dataset Characteristics
The Power BI report incorporates datasets that encompass a range of metrics, including COVID-19 cases, deaths,
hospitalizations, ICU admissions, ventilator usage, and administered vaccination doses. These datasets cover all
states of Australia, namely New South Wales, Victoria, Queensland, South Australia, Western Australia, Tasmania,
Northern Territory, and Australian Capital Territory.

2.1. Data Fields/Variables

The column headers within these datasets have been modified to improve their clarity and relevance for the target
audience. A comprehensive explanation of these modifications is available in Section 3.2 of the report.

The data fields/variables in Covid-19 dataset provided include:

Date: the date on which the COVID-19 data was recorded.

Covid Cases: the number of confirmed COVID-19 cases.

Covid Deaths: the number of COVID-19 deaths.

Covid Cases Hospitalised: the number of COVID-19 cases that required hospitalisation.

Covid Cases ICU: the number of COVID-19 cases that required admission to an intensive care unit.

Covid Cases VENT: the number of COVID-19 cases that required the use of ventilator.

Covid Vaccination Doses: the number of COVID-19 vaccination doses administered.

State Code: the code representing the State in Australia where the data was recorded.

The data fields/variables in States dataset provided include:

State Code: this field represents the unique code assigned to each state in Australia.

State Name: this field represents the official full name of each state in Australia.

State: this field represents the abbreviated from of the state name.
2.2 Data Type:
The appropriate data type for each of the column headers above are:

Date: Date

Covid Cases: Whole Number

Covid Death: Whole Number

Covid Cases Hospitalised: Whole Number

Covid Cases ICU: Whole Number

Covid Cases Vent: Whole Number

Covid Vaccination Doses: Whole Number

State Code: Whole Number

State Name: Text

State: Text

3. Data Cleaning and Formatting

Data cleaning and formatting play a vital role in enhancing data quality, minimizing errors and discrepancies, and
improving the efficiency and accuracy of data analysis. In Power BI, several data transformation techniques were
utilized to clean and format the two datasets effectively. These techniques included actions such as promoting
headers and adjusting data types, renaming columns, adding and merging columns, and replacing values, among
others.

3.1. Promoted headers and changed data type

In Power BI data preparation, promoting headers and adjusting data types are crucial steps that ensure data
consistency, accuracy, and relevance, ultimately enhancing data comprehension and usability. Defining data types
accurately can also improve query performance, simplifying tasks such as aggregations, filtering, and sorting. In this
particular case, the "Use first row as Header" functionality within Power Query Editor was employed to promote
headers.
3.2. Renamed columns
The objective of renaming columns is to enhance their descriptiveness, relevance, and understandability for the
intended audience. Renaming columns aids in identifying and resolving data quality issues, such as inconsistent data
types or missing values. For example, if multiple columns contain similar information but have different names,
renaming them simplifies the identification and consolidation of redundant data. In order to improve the readability
and precision of the datasets within Power BI, the following modifications were made to the column headers:

The columns labelled 'New Covid Cases', 'New Covid Death', and 'New Covid Vaccination Doses' were renamed by
removing the word 'New'. This adjustment aligns the column names with the naming conventions used for other
columns, enhancing the clarity of the data.

'Territory key' was changed to 'State Code' to create more accurate and standardized column headers. Since this
column contains the same data as the 'State Code' column in the State dataset, establishing a relationship between
these two datasets is possible using this column.

‘Official Code State’ was renamed to ‘State Code’ as a similar column header and the data it contains exists in the
Covid-19 dataset.

‘Official Name State’ was renamed to ‘State Name’

‘Iso 3166-2 Feature Code’ was renamed to ‘State’.

3.3. Adding and merging columns:

Including additional columns is a significant approach when it comes to cleaning and formatting data in Power Query,
as it enables performing calculations and transformations that simplify data analysis. In my process, I utilized the 'Add
Column' tab within Power Query Editor to incorporate new columns. This feature facilitated various operations, such
as column splitting, merging, and custom calculations. Notably, adding columns like Year, Month Name, Day Name,
Month, and Week of Year to my Power BI data proved invaluable in streamlining data analysis and visualization. By
incorporating these columns, I seamlessly organized and summarized data based on specific time periods, enabling
the identification of trends and patterns over time using the powerful and user-friendly "Visualizations" pane.

The process of merging columns plays a vital role in ensuring data consistency and accuracy. When separate columns
exist for the year and month, combining them into a single column helps eliminate errors and inconsistencies that
may arise if users have to manually enter this information.

To address this, I merged the Year and Month Name columns, resulting in a new column named 'Year-Month Name'
that displays values like '2022-January'. Similarly, by merging the Year and Month Number columns, I created another
column that presents values such as '2022-1'. This approach has proven particularly beneficial when sorting and
filtering data based on specific time periods and conducting calculations using the data.
3.4. Replaced Values
The process of replacing values within a dataset is an essential aspect of data cleaning and formatting that serves
multiple purposes. Values such as 'n/a' or blank cells can lead to inaccuracies and errors during data analysis,
potentially introducing biases if not dealt with properly.

By substituting 'n/a' values with appropriate alternatives, such as 0 or the mean of the column, the dataset becomes
more consistent, complete, and accurate, enabling precise analysis and informed decision-making. In my case, I have
replaced all 'n/a' values with 0, ensuring the dataset's consistency and completeness.

4. Establishing Relationship between Data Tables

Creating relationships between data tables is a pivotal step in organizing data in a structured and meaningful way,
which is crucial for efficient and effective data analysis. Establishing relationships allows for the integration of data
from various sources, enabling cohesive analysis.

In this scenario, we have a fact table and a dimension table, and the relationship between them is established using
the 'State Code' column. Analysing Covid-19 data by state provides several advantages, including the reduction of
data redundancy and errors. It ensures that any changes made to one table are promptly and accurately reflected in
all other tables that share the same data. By establishing relationships, data integrity is maintained, enabling reliable
and comprehensive analysis of the Covid-19 data at the state level.
5. Report Pages
In Power BI, the creation of a report page empowers you to present and visualize data from diverse sources in a
coherent and meaningful manner. By leveraging report pages, you can analyse and convey insights with clarity and
effectiveness. These pages provide a platform to transform raw data into visually captivating charts, graphs, tables,
and other visual representations. This aids in comprehending patterns, trends, and relationships embedded within
the data. Moreover, report pages allow for the incorporation of interactive elements such as slicers, filters, and drill-
through actions. This interactive functionality empowers users to dynamically explore the data, filter it based on
specific criteria, and delve deeper into the intricate details. Ultimately, report pages in Power BI serve as a powerful
tool for data exploration, analysis, and communication.

5.1 Visuals and the reason behind their choice

In order to convey the findings obtained from data analysis in a compelling manner, the Power BI report incorporates
a diverse array of visual elements. These include, but are not limited to, cards, slicers, line and stacked column charts,
clustered bar charts, maps, donut charts, line charts, area charts, clustered column charts, and treemaps. The
selection of these visuals within the Power BI report was influenced by various factors, which will be elaborated upon
in the following discussion.

Cards have been utilized in the Power BI report to present important summary information in a concise manner. A
variety of cards have been incorporated, each specifically designed to showcase key insights regarding the data.
These cards encompass various metrics such as the total number of covid cases, deaths, hospitalizations, ICU cases,
ventilator cases, vaccination doses administered, and more. This comprehensive selection of cards enables users to
obtain a holistic understanding of the overall situation at a glance, providing valuable insights into the data.

The inclusion of cards was based on their ability to present succinct summary information, allowing viewers to
quickly grasp key metrics. For instance, Australia has recorded a total of 10,724,217 cases of Covid-19 over a span of
three years, while also administering a significant number of vaccinations to combat the virus's spread. These cards
highlight Australia's proactive approach in reducing the number of cases and preventing further outbreaks.
Additionally, the data reveals that over one million individuals were hospitalized due to Covid-19, underscoring the
profound impact of the virus across all states of Australia.

Slicer is an influential data visualization tool within Power BI that empowers users with interactive filtering and data
selection capabilities. With slicers, users can effortlessly filter data within a report or dashboard, directing their focus
towards specific subsets of information and deriving insights accordingly.

In Power BI, slicers significantly enhance the interactive and exploratory aspects of reports, enabling users to
dynamically filter, analyse, and compare data subsets. By providing a user-friendly means of controlling the displayed
data, slicers elevate the usability and effectiveness of Power BI reports.

The combination of a line and stacked column chart in Power BI offers a robust approach to data representation and
comparison. This chart allows for comparative analysis by presenting two measures, with one represented as a
stacked column section and the other as a line graph. By leveraging this chart, users can easily compare trends,
variances, and relationships between the measures.
The primary reason for selecting this visual was to showcase the total number of Covid-19 cases for each state,
enabling viewers to compare the overall number of cases across states and identify trends in the data over time. The
line chart component within this visual effectively highlights these trends, particularly the decline in cases for each
state.

The visual representation provides crucial insights into the total number of cases, hospitalizations, VENT cases, and
ICU cases. It reveals that approximately half of the hospitalization cases required the use of a VENT.

The Donut Chart offers a succinct and visually captivating method to showcase proportions and percentages of data.
This visual representation effectively displays the percentage of Covid deaths by year, presenting the relative
proportions of deaths in a visually engaging and intuitive format.
The Clustered Bar Chart is a powerful tool for comparing data across different categories, facilitating the
identification of patterns with ease. Its flexibility and customizability make it valuable for data analysis and decision-
making purposes. This visual is particularly well-suited for visualizing the total Covid-19 cases by month, as it
effectively compares data across different time periods, enabling the straightforward identification of trends and
patterns within the data. As depicted in this chart, January exhibited the highest number of cases.

Maps present data in a visually captivating manner and provide valuable insights based on location. In this visual,
Covid-19 deaths are represented by state, offering a practical and intuitive approach to showcase the geographic
distribution of the data. The size of the bubbles accurately reflects the number of Covid-19 cases, enabling the
identification of regional or spatial patterns within the data. This information holds significance for informing public
health policies and interventions, as it provides insights into the areas most impacted by the pandemic.

Line charts are a visual representation that showcases data points connected by straight lines, commonly used to
depict trends over time and compare data within a dataset.
In this scenario, the objective was to examine the relationship between the number of Covid cases and deaths on a
monthly basis. Throughout the three-year period, New South Wales recorded the highest number of Covid cases
nationwide, totalling 886,343 cases, while Victoria experienced the highest number of deaths.

The second visual, represented by the line chart displaying Covid vaccination doses by month, illustrates the changes
in the number of administered vaccination doses over time. In the year 2020, no Covid vaccinations were recorded.
However, in the subsequent year, a consistent increase in the number of doses administered can be observed.
Interestingly, there was a notable decline in the number of vaccinations administered during the year 2022.

Area Charts, like line charts, visualize data trends over time. The filled-in area in an area chart serves to accentuate
the magnitude of changes and draw attention to significant increases or decreases within the data.

In this specific chart, the focus is on capturing the viewer's attention and highlighting Covid-19 hospitalizations,
including those in the ICU and on ventilators. It becomes apparent that a majority of hospitalized patients required
intensive care and mechanical ventilation. Moreover, a noteworthy surge in hospitalizations, particularly in the ICU or
with ventilator usage, is observed starting from August 2021 and continuing onward.
100% stacked bar charts allow for the depiction of the relative proportions of different categories as percentages of a
whole. This choice was made to facilitate the comparison of the relative proportions of Covid-19 hospitalization
cases, ICU patients, and the percentage of patients on ventilators.

As observed, the highest number of cases were those requiring hospitalization, while the percentage of patients on
ventilators was relatively low. This visualization proves to be particularly valuable in showcasing the relative sizes of
each category, thus enabling easier comparisons between them.

Treemaps are powerful tools for representing complex data in a visually appealing and engaging manner, while also
providing a clear visual representation of data hierarchy. In this context, a treemap was selected as the ideal visual to
effectively display the hierarchical distribution of Covid-19 vaccination doses across different states in Australia.

The treemap reveals that New South Wales (NSW) had the highest total of 19,944,209 doses, making it the state with
the most vaccinations. Victoria, Queensland, and other states followed in descending order, each with their
respective vaccination totals. This treemap visualization allows for easy comparison and understanding of the
distribution of vaccination doses among the states.
The Clustered Column Charts serve the purpose of comparing data across different categories. It shares similarities
with a standard column chart but distinguishes itself by allowing the display of multiple data series side-by-side,
facilitating easy comparison.

The selection of this visual was motivated by the intention to showcase the impact of vaccination rates on Covid-19
cases. Through this visual, it becomes apparent that a substantial number of vaccination doses were administered
across all states, with a remarkable 8,690,068 doses administered in September alone. This observation signifies a
proactive approach in controlling the spread of the virus and serves as a positive indication of the effectiveness of
vaccination efforts in combating the ongoing pandemic.

5.2. Applying Visualisation Formatting Option

The mentioned visuals have undergone various customization techniques to improve their visual appeal and
readability. These formatting options include color-coding segments to represent individual states, adding axis labels
and titles, and adjusting the chart layout to effectively display the data. These customization techniques have been
applied to enhance the overall visual presentation and ensure that the information is easily understandable to the
audience.
6. Applying Filters for Interactive and Data Narrowing
The Power BI report integrates a variety of filters to provide users with the ability to interact with and explore the
data in diverse ways. These filters empower users to narrow down their focus to specific data points and subsets that
are pertinent to their analysis. By leveraging these filters, users can refine their view of the data and extract valuable
insights tailored to their specific needs.

6.1. Applying Filters at Different Levels of the Report

Filters are implemented at various levels within the report to offer users a comprehensive and flexible experience.
Visual filters enable users to narrow down data within specific visualizations, while page and report filters allow for
filtering specific pages or the entire report. Additionally, drill-through filters enable users to explore related data in
other visuals based on their selections.

By incorporating filters at different levels throughout the report, users can refine the data and delve into specific
trends or patterns, making the report a powerful tool for analysis and decision-making.

In the Power BI report, I have applied visual filters that allow users to filter the data by date, year, month, and state,
providing insights into the impact of the pandemic. Notably, on the first page of the report, a report filter has been
set up using the 'Year' column, enabling users to filter all pages based on a specific year. On page 4, 'Medical State', a
page-level filter has been applied using the 'State Name' column, while on page 5, 'Vaccination Progress', a drill-
through filter has been implemented using the 'State Code' column.

7. Conclusion
In conclusion, this report exemplifies the remarkable capabilities of Power BI as an impactful data visualization tool.
Through a thorough analysis of the Covid-19 situation in Australia and its states, it has effectively demonstrated the
ability of Power BI to extract valuable insights and identify trends from intricate data sets. By addressing key aspects
of Power BI, such as data cleaning, formatting, visual selection, and the utilization of filters, the report has
successfully achieved its objectives and stands as a testament to the strength of Power BI as a robust tool for data
analysis and visualization.