We have previously talked about the best practices in Data Visualization and the art of telling a story with it. Now it is time to dive into the detail. Starting from choosing the most suitable tool for your visualization, we will share some tips and tricks on how to improve your data presentation and convey the meaning of your data analysis.
If you are here only for the white paper, scroll all the way down. The article is great too though. 🙂
All 3 tools have the “basic” free versions and the “advanced” paid versions. The key differences between them are ease of operating and cost-effectiveness. Power BI wins the race for ease of operating, because of the Microsoft Excel background and the general widespread usability. As for the cost-effectiveness, Power BI is the least expensive of the 3, but the choice here should account for functionality.
Tableau in its basic version is free with limited features. The advanced version costs $100 per user and provides the best analytics and interpretation option out of the 3 tools, due to the many drill-down and filtering options. Tableau has good data source connections if your focus lies in Data Capturing and Warehousing.
PowerBI’s free version includes the majority of the features and users can download the software and start using it right away. There are 2 advanced versions: Power BI Pro & Power BI Premium. These packages cost $10 per user per month, which is less expensive than Tableau and Qlikview. Power BI has a lot of advanced features to work with analytics and interpretation of the data. PowerBI has the capability to fetch data from any source (web, cloud SQL, Azure, etc.), which is helpful for Data Capturing and Warehousing work.
Qlikview has a limited number of features in the free version. The paid plan is $30 per month per user. Unfortunately, Qlikview lacks the same analytics and interpretation features as the other two, which may be the reason for its limited popularity. Data transformation is the key element of Qlikview and this tool too can fetch data from anywhere.
There is a lot to consider when choosing the preferred visualization tool. Here are some questions to help choose wisely.
1. Do the chart and graph types meet the business needs to show the right type of analysis?
Does the tool allow chart extensions or custom charts? You need to consider the business context of the visualization tool under scrutiny. The best tools are flexible tools, and this is hard to achieve. Rather, you should consider what the needs of your business are and how your data and results need to be visualized and output into a report.
Then, make sure that your candidate tool can accommodate this. This is a crucial step in selecting a good visualization tool, since choosing the wrong tool and discovering later on that it does not meet your business needs can lead to a lot of rework needed and time wasted.
2. What are the data transformations that should be supported by my tool of choice?
Many data visualization tools offer the ability to make back end transformations on the data. Watch out for the ones that don’t. They are not worse, but make sure that you do not need the transformation functionality.
Some examples of data transformations that may be very useful to have are simple group-by summaries, pivots and unpivots, data type manipulations, and string manipulations. Often any transformations that can be done inside the tool of choice are a time-saver versus maintaining an external processing pipeline. Especially, if these in-house data transformations can be automated for future-proofing of your visualizations.
3. Are the necessary data connectors available in the tool?
The data connectors are required to connect to the right data sources from within the tool. If your tool can connect to a wide array of data sources, from CSV, Excel, to live sources of data such as SQL databases of various flavours and beyond, then you’ll find you have a lot more freedom to accomplish more with your data visualization tool across projects.
4. How large are your datasets?
Some tools have a limitation on the number of rows of data that can be loaded. And while some tools might not have such a limit, you may find certain tools become progressively slower and less seamless in their visualizations as you increase the size and complexity of your data.
Good tools are more scalable and maintain their ease-of-use and user-friendliness as much as possible. However, it is also up to the user of said tool to ensure that the data transformations applied to the data are performed in an efficient way and that the data provided is the minimally required data for visualization purposes.
5. Do the reports need to be automated?
Some tools have most of the necessary connectors to connect to live data, and then schedule the reload and distribute jobs to enable full automation. The real question is whether such a job can be automated via the tool under consideration. Do you need reports to be generated on a frequent basis? Can this report be exported in a specified format, or sent via email to relevant stakeholders? Is the tool able to do this with this flexibility in mind?
6. Who needs the ability to create reports?
The tool must have the provision of self-service capability if required. In conjunction with the previous point, users should not only be able to generate periodic reports but also to generate on-demand reports. Another great tool capability is to customize the output in a report, be this special views of existing charts with specific filters, or only a subset of the information in the full dashboard.
7. What devices and layouts does the tool support?
Some tools can only be used on devices with one specific type of layout and resolution. If the report needs to be accessed through multiple devices, make sure the tool has this functionality.
Another important factor is the layout capabilities. You should choose a tool that not only has expressive visualization power but also allows you to customize the visual layout of charts and graphics within the report or dashboard. Having this capability in your tool allows you to easily and seamlessly create a visual story using the charts you have built to paint a picture of your data.
8. Is the security configurable for the required end-user access rules?
In most cases, the visual report needs authentication and authorization of different levels to enable different individuals or teams to see certain parts of the data and not all of it.
Many tools have in-built security mechanisms, while others rely on an organization’s policy regarding account access rights, and some rely on security to be provided by the hosting service where the reporting will be deployed. Choose which type of security offering meets your business needs best, making sure it serves your end-users in the most user-friendly and accessible way possible, while, of course, keeping the data secure.
There may be more than one way to visualize the data accurately. Consider what you are trying to achieve, the message you are communicating, who you are trying to reach, etc.
Refer to our Expert Guide to Data Visualization article to choose the right chart for the business purpose.
- Remove the dark chart border.
- Eliminate the gridlines.
- Use data labels where appropriate.
- Thicken the bars to reduce the white space in between.
- Ensure that the axis labels are appropriate.
- Avoid diagonal labels when possible.
- Place the legend relative to the data it describes.
- Set the data labels to whole numbers.
- Use colors wisely, red usually is a negative indicator and green is positive, this could cause confusion and take the focus out of the actual point.
- Always add a descriptive title. It ensures that the user clearly understands what you are trying to convey with the visual.
- Draw attention to the necessary data points.
Pro tip: never use 3D charts as it causes ambiguity in reading the data points due to the depth and 3D visual effect. It might cause the wrong data interpretations.
Numbers and charts do a great job of filling our rational need for quantifiable information. But when it comes to communicating how things will impact our real lives, or affect the world we live in, some form of humanizing or grounding the data is often effective.
Use softer color tones except where you want to draw attention. Distinguishing the colors of the data points is a great way to draw attention to them. Here, how fast can you spot the data points that have not met the target?
Callouts are not there to fill space. They are used to highlight relevant information or provide additional context.
Here, more data for other channels is expected at a later stage. The callout saying that the “Data currently includes only GMB reviews” adds relevant information for accurate analysis.
If you are depicting sales month by month on a bar chart, use a single color. But if you are comparing last year’s sales to this year’s sales in a grouped chart, use a different color for each year. Make sure the colors are sufficiently contrasted and not too similar in shade.
Use a different color when you need to highlight specific data points.
Use green for positive data points and red for the negative ones. Using green and red in a visual that does not necessarily show a positive or negative impact can be misleading – don’t.
Do not use patterns like polka dots, stripes, etc., instead of colors.
Do not use too many colors in a single layout.
Follow these tips to build effective and neat visuals that will help the user engage with the report better. To make it easier for you, we have written a white paper on data visualization and storytelling, which you can access by completing the form below. The white paper is divided into 8 sections and includes everything from rules and guidelines to do’s and don’ts and some pro tips. All yours, for free!