The majority of objects in QlikView are capable of handling several dimensions and expressions without any trouble. Unfortunately, we cannot say the same about our brain. When dealing with numerous items cluttered in the same area, it is quite easy for us to lose focus, so try to limit the amount of data you display in a single chart.

Most of the time, keeping the number of dimensions and expressions under four will work fine (1 expression and 1 dimension, 2 dimensions and 1 expression, 1 dimensions and 2 expressions, and so on). In this regard, there are three notable exceptions. The first one refers to combo charts. Since they work with different representations at once (bar, lines, and symbols), they can handle more expressions than the average chart. In contrast, pivot tables only make sense while using multiple dimensions as each item adds a new level to the hierarchy. Finally, pie charts are only useful while working with one dimension and one expression, so they are better kept simple.

Some visualizations—such as top 10 rankings—work well in either orientation. Most of the time, the decision to use a vertical or a horizontal graph has more to do with style than functionality. However, horizontal representations are often preferred because they can display long labels clearly, and you can add the values at the end of the bars without the fear of overlapping numbers.

Conversely, there are some scenarios where the orientation does affect the chart's readability. For example, when representing a trend over time, it is better to stick to the classic horizontal orientation because our brain is accustomed to reading from left to right, so it is natural for us to interpret the oldest point on the left-hand side and the newest elements on the right-hand side.

Depending on the context of the data, it may be beneficial to remove the axes lines in order to maximize the data-ink ratio and present a leaner design to the users. For instance, the following graph already shows the values at the end of each bar, so it may be redundant to include a scale on the axis as well. In addition, the gray line between the labels and the bars doesn't add any value to the visualization, so we can delete it without remorse:

In the same manner, remember that intense colors are to be reserved for the most important elements of a chart. If you use lighter tones on axes lines, grids, and labels, your visualizations will benefit at both the aesthetic and functional ends.

It is always important to consider whether the axes will have an impact on the interpretation of a chart. If your main focus is to show a comparison of magnitudes, you should force the axis to zero (Axes tab). This option will display each element as big as it is and will usually make your charts look steadier. On the other hand, if you need to focus on the variances amongst the items—even if they are small in comparison to the total magnitude—it is better to let QlikView adapt the scale so that the differences are more evident.

Forcing the axis to zero in a bar or line chart is neither good nor bad; it all depends on what you want to highlight. Just watch out for charts like this one (yes, I am using a 9GAG post as an example):

Occasionally, in-chart selections are undesirable due to the nature of the expressions or the dashboard's navigation schema. If you need to disable the selections in a specific object, go to the General tab and activate the Read Only box.

Cluttered charts are rarely useful, so be careful when working with high cardinality dimensions (fields with many distinct values). Most of the time, it is better to limit the number of elements displayed for readability's sake. If it is enough to show only the top (or bottom) items, you can work with the Dimension Limits tab. On the other hand, if it is valuable for the analysis to see all the elements, it is better to add a scrollbar in the Presentation tab.

Again, it all depends on your visualization's objective. Generally, grouped charts foster the comparison between individual items, while stacked representations highlight the composition of the bars. In either case, you should be careful while dealing with high cardinality dimensions as your chart might turn into an unreadable stained glass.

Don't be afraid of changing old-fashioned charts with innovative alternatives that help the user better visualize the data. For example, using a bar chart with two expressions is a common way of comparing KPIs. However, if both calculations are too similar, it can be hard to see the variances over time. Take, for instance, the following graph representing the new credits positioned by a bank against its budget:

As you can see, the magnitudes are so big (and the gaps between the metrics are so small) that it is difficult to draw conclusions in this regard. If your goal is to highlight the small differences between the actual and expected figures, you can change the perspective of the visualization and plot the variances instead of the actual numbers:

Actually, mixing both objects (maybe using a container) will give the user a better perspective of the situation. In this way, they will be able to see trends in the global magnitudes but also dive into the details and make precise comparisons.

If you research data visualization, you will soon realize that there are plenty of detractors of pie charts. While it is true that they have some downsides, a good designer knows how and when to use each tool, so do not throw them away just yet.

Pie charts are mainly used to represent parts of a whole, so they work best with percentages. However, as our brain is not very good at measuring areas, sometimes it is hard to distinguish which slices are bigger without the help of a legend (sorting the values can help too!). Hence, the pie chart is a bad choice for situations that require a detailed comparison.

Likewise, avoid this visualization with high cardinality dimensions. In other words, don't use it if you will end up with too many slices. If you are dealing with several small items, it is better to restrict the number of values shown by including an Others slice (Dimension Limits tab).

One of the hidden charms of pies is their shape. While lines usually require a horizontal, rectangular area and bar charts demand an elongated space, pie charts can adapt a little better to difficult spaces.

Albeit that it is a little difficult to master this type of chart, they are quite popular amongst business users, so it may be nearly impossible to get rid of them. Remember that there are no bad tools, only misapplications (well, maybe the radar chart IS a bad tool, but you get the point).

Have you ever printed a QlikView object using the caption icon? Well, neither has the vast majority of users around the globe. Even though all charts include multiple icons in the caption bar by default, it is better to include only the relevant ones. By doing this, you will also reduce the chartjunk and create a leaner visualization. In this regard, don't forget that most of these functions are available through the contextual menu (that is, by right-clicking it); so, it is possible to hide most of the icons without reducing the object's functionality.

When used adequately, color can greatly improve your visualization by making it easier to read or adding valuable information. There are several ways of taking advantage of colors within a chart, for instance:

Time to talk about one of the most relevant parts of a chart: the legends. The correct usage of labels boosts the usability of any dashboard by adding context to each visualization (the type of information being displayed, units of measurement, or the base of comparison, just to name a few). There are several ways of labeling a chart, and it is your job as a QlikView designer to decide which ones to include, their position, and their format.

These are the most common ways to include labels inside a chart: