Data visualization is the representation of data using common graphics like charts, plots, infographics, and even animations in an easy-to-understand manner.
In Power BI Desktop, different visuals are made available and can be found in the Visualization pane. Others can be enabled via preview features. File >> Options and settings >> Options >> Preview features. You can also get more visuals by clicking on the three dots as seen below.
Choosing what type of chart to use is a key component of a data visualization strategy. A great tip is to consider which one best conveys the story and insight of your data.
Visuals in Power BI
1. The Bar Chart (Stacked and clustered bar charts)
Bars charts are one of the most common charts used in data visualization. As long as there aren’t too many categories to compare, bar charts can be useful for comparing various categorical or discrete data, such as age groups, classes, schools, etc. Positive and negative values can be displayed on the axis of the bar chart. They accommodate longer axis labels and the bars are horizontal. Bar charts could be clustered or stacked to show the composition of variables.
Design best practices for bar charts
- Always use consistent colours.
- Make sure the y-axis starts from 0.
- Avoid displaying too many bars to improve readability.
- Ensure that the bars are sorted in descending or ascending order.
2. Column Charts (Stacked and clustered Column charts)
Interestingly, bar charts and column charts are most times seen as the same. Although they are similar, the major distinction between both charts is in the usage of categorical labels. Long labels don’t suit column charts because they may become diagonal or reversed vertically which makes them difficult to read. Column charts work better with time series data as well as categorical data. Column charts could also be clustered or stacked to show the composition of variables.
Design Best practices for column chart
- Avoid labels that are vertical or diagonal. Always use horizontal labels to assist readability.
- Make sure the y-axis starts from 0.
- Always use consistent colours.
clustered column chart
3. Line charts (area and stacked area)
Line charts are most suitable for time series data, to measure changes over a period of time. In Power BI, line charts come in 3 forms – basic line charts, area charts and also stacked area charts. Area charts are usually based on a line chart, colours are used to indicate volume in the space between the axis and line, and lines depict trends over time. Area charts may be stacked, dividing the data into sections that represent part of a whole or cumulative data.
Design Best practices for line and area charts
- Ensure you use horizontal labels for better readability
- Avoid dotted lines, always use solid lines
- The fewer the lines the better
- For area charts, transparent colours are mostly preferred
- When plotting more than one line, always use legends
- Do not use more than 4 lines when using area charts to avoid misinterpretation of the colour shades.
4. Waterfall Chart
A waterfall chart demonstrates how the cumulative effect of sequential positive and negative values can affect an initial value. Using this graph, you can display categorical or sequential data. It is often used to represent changes in financial results, such as revenue, expenses, profit or net income over a period of time.
Design Best practices for line Waterfall Charts
- Use contrasting colours to highlight differences in data sets.
- Choose warm colours to indicate increases and cool colours to indicate decreases.
In Power BI you can create four types of map charts, generally, map charts work best with geographical data. For example, the map chart can be used to compare sales performance across regions.
- Basic Map: Usually used to plot the numeric distribution of data in the form of data points. It leverages Bing Maps.
- Filled Map: Plots data points as geospatial areas rather than points on a map. They usually are used to plot the geographical distribution of categorical data.
- Shape Map: The Shape Map visual is only available in Power BI Desktop. Since it’s in preview mode, you’ll have to enable it before using it. These plots are based on Esri/TopoJSON maps and can be used to plot Power BI custom map visuals as well.
- ArcGIS Map: ArcGIS for Power BI uses a powerful geo-enabling technology to precisely place location data on the map depending on the type of data and by default displays things as points or boundaries. It allows you to plot maps containing street views and heat maps, with advanced customisation options.
Note: If you are using a map for the first time on Power BI desktop you will have to enable it from the security settings.
The steps listed below will help enable Power BI Maps:
- From your Power BI Desktop click on the file at the top left corner, this should open a new tab. Then select “Options & Settings” then options.
- In the Options, select the security tab, this should open a new environment, then scroll down to see the “Map and Filled Map Visuals” and then enable it.
- Click on OK when you are done with the setting. You are expected to close your Power BI Desktop and re-open it to activate the configuration setting you just created.
Cards are used to display a single value or metric, usually accompanied by a label or title. They are also useful for displaying key performance metrics and summarised data. Power BI offers the Multirow card that helps to summarise more than a single value or matrix once in just one card.
Design best practices for Card Charts
- Avoid too many cards while creating a report.
- Make sure labels are clear and visible.
6. Pie and Doughnut Chart
Slices are used in a pie chart to compare different portions or categories. The only difference between a doughnut chart and a pie chart is that a doughnut chart has a hole in the middle. Pie charts and doughnut charts are best used for comparing at most 3 (three) categories of data.
doughnut chartpie chart
Design best practices for Doughnut and Pie Charts
- Avoid using piecharts if the difference between categories is not so significant. Opt for a different chart type such as a bar or column chart.
- Always go for 2D effects and Avoid 3D effects as they may distort the data and make it more difficult to read.
- Consider placing the data labels outside the chart or opt for a legend if the space is limited
7. Guage Chart
Gauge charts are often used to display KPIs, such as the percentage of a sales target that has been reached, or other metrics critical to a business like net promoter score, customer effort score etc. They are used to demonstrate advancement towards a specific objective. It aids in determining the degree of accomplishment of a task. Key performance indicators (KPIs) like sales, income, personnel productivity, or profitability are frequently represented using gauge charts.
Design best practices for Guage Charts
- Ensure you use colours with a logical meaning behind them. For example red for a poor performance and green for a great performance.
- For measurements, ensure the measurements start from Zero
- Always make sure the data labels are legible enough to be seen at a glance.
8. Decomposition tree
It is a hierarchical representation of data that shows how a single metric is decomposed into smaller, more granular components. It is a fantastic drill-down feature that can help with root-cause analysis. they can help to break down large data sets into smaller, more manageable pieces, making it easier to identify trends and patterns.
Design best practices for the Decomposition tree
- The decomposition tree displays data hierarchically, choose the levels of detail that are appropriate for your analysis
- Make sure to customise the decomposition tree with minimal colours in a way that is visually appealing.
9. Tables and Matrix
Tables and Matrices are like mini spreadsheets that display information in rows and columns. Although the two visuals are similar, they have striking differences in their use case. A matrix functionality is similar to pivot tables in excel and allows for some form of drill-down and hierarchical functionalities. Tables don’t have drill-down and hierarchical functionalities.
Although other visualization charts are excellent options for simplifying complex data sets, tables and matrices are ideal when you have data that cannot be easily presented visually, because they are designed to be read. Visual elements like arrows and colours could also be added to tables using the conditional formatting option.
Design best practices for Tables and Matrix
- Avoid too many columns and rows. If the table has more than ten to fifteen rows and five columns it becomes difficult to read, understand, and gain insight from.
Power BI slicers are used as graphical filters within a report which makes the report very interactive for the end users. Slicers could be lists, dropdowns, between or even tiles. They are most commonly used as date filters.
Design best practices for Slicers
- Avoid too many slicers, only use slicers when necessary
11. Tornado chart
This is one of the most commonly used custom charts. They are a great choice for comparing variables in a data set. For example, you can compare the number of units sold versus the total amount of revenue from each regional market.
Design Best practices for Tornado Charts
- Make sure to use clear data labels
- Make sure the variables being compared have a logical interpretation
12. Funnel Chart
A funnel chart is primarily used to illustrate a sequential process from top to bottom. The data set at the top of the process is typically larger than the one at the bottom so the quantity decreases as it flows down.
Design Best practices for Funnel Charts
- Use contrasting colours or one colour with adjusted shades from darkest to lightest as the funnel size decreases.
- Adjust the size of each section to accurately reflect the size of your dataset.
13. Scatter Plot
The chart displays points at the intersection of numeric x and y values and combines those values into a single data point. These data points can be evenly or unevenly spaced on the horizontal axis, depending on the data. The scatter plot is commonly used for identifying the relationship between two numerical variables usually on large datasets. The closer the dots usually represents a correlation between the two variables which may be a negative or positive correlation.
Best practices when using a Scatter plot
- Always start the Y axis with 0
- Avoid using more than two different dot colours and always use a legend when using more than one dot
- Do not plot too many data points as it may become impossible to read
- Use the scatter plot for continuous data only.
14. Tree Map
A tree map uses coloured rectangles to display data. Space within each rectangle is allocated based on measurements. Rectangles are sized from top left (largest) to bottom right (smallest).
Design best practices when using Tree Map
- Avoid including too many composite values. If you have a lot of data, it’s advisable to use another value
- Avoid too many contrasting colours
- Make sure labels are clear and visible.
- Choose hierarchical data over numerical data when working with treemaps
15. Word cloud
Word clouds are an easy-to-use custom visual in Power BI usually used for text visuals. It is a collection or cluster of words that are depicted based on size. The larger and bolder the word appears in the cloud, the more it was mentioned within a text or speech. However, one of the challenges in interpreting word clouds is that the display emphasizes word frequency, not necessarily word importance.
Best practices for using a Word Cloud Visual
- Avoid stop words, stop words are words that appear very frequently but do not convey any meaning to the context. Examples are words like and, at, on, from etc.
16. Ribbon Chart
The Ribbon chart is similar to a stacked column chart but instead, ribbons connect columns to each other and also show a flow. Ribbon Charts take up to 3 variables to be measured, ie. X-axis, Y- axis and the Legend. Without the legend, the column would have no connections and would look like a column chart. Ribbon charts are effective for identifying and ranking data in categories. It shows the higher value in each column at the top, then the next value comes after. They can be particularly effective for showing the changing patterns of multiple categories over time.
Design Best practices for Ribbon Charts
- Avoid the use of too many categories as they can clutter the chart and make it difficult to read.
- Use the ribbon chart for categorical data rather than continuous data
Rounding Things Up
In summary, the type of chart you choose should depend on the question you are trying to answer while analyzing data. Below is a cheat sheet to show the most appropriate charts to use depending on the scenario.