10 Ideas Every Professional Should Avoid for Data Visualization – News Couple
ANALYTICS

10 Ideas Every Professional Should Avoid for Data Visualization

Data is the new oil we’ve heard so many times. But can we visualize that data in a proper format? With the amount of data we have today, the need to extract ideas from it is more important than ever. Hundreds of visuals are created every day. Some of them are well appreciated by the public while others are rejected. Why is that? Well, the answer lies in creation. Let’s find out the cause and problem and see how to solve it.

Here, I will summarize some of the best and worst versions of the charts, so you can stop this if you do.

The graph starts with the baseline 0

One of the most common mistakes I’ve found while plotting is not starting the graph with a baseline of 0 and using some random values.

Use the correct tape chart

We have commonly used horizontal or vertical bar charts to visualize data. Sometimes when we use simple bar charts to compare well it conveys the message but stacked bar charts in vertical or horizontal stacked bar charts is better. Let’s see with an example.

Example 1:

We have monthly sales data for females and males. Both charts show a comparison of the sex ratio for each quarter. We can use the chart below for comparison.

Never plot positive and negative values ​​on the same side of a chart or never plot to compare features on the same side of a chart, as the chart becomes difficult to read.

Multiple colors on the chart

Multiple colors must be used for some reason in the graph. The silly use of colors in a graph is a real turn off while seeing the graphs. Mostly only two colors are used in charts.

If you have more than two charts, it would be a great idea to separate the charts by colors. Let’s see how we can do that.

As we can see here each column has its own purpose so giving the same colors wouldn’t look very interesting that’s why here I’ve given the same colors to charts that represent the same purpose only.

Data confirmation

Sometimes stressing data also makes charts look pretty. At first glance, we will get a file NS the value higher Among all the other features we used.

Here we only have 4 features so this simple graph change will not be effective but when we work with 100 features highlighting the highest feature value will help us a lot. Getting and planning 100 features is a difficult task but at that moment we can pass a condition of suppose the top 10 features will be highlighted in gray while others will be pink.

Baffling choice of colors

Lines, colors, and axes are all important when drawing diagrams. Choosing chart colors is a very important step because if you have very bright or very light colors on the charts, the charts will be difficult to read.

In the example, we have two different charts showing highest to lowest sales in an area on colours. If we see the graph on the left side, then with the naked eye no different shades of yellow will appear. On the right side, we have a blue and pink color scale where we can easily distinguish between shades.

Avoid randomness on charts

Always put the bars in ascending or descending order according to their values. Put the largest value at the top for horizontal bar charts and place the largest values ​​to the left for vertical bar charts. This will help the audience to determine the highest and lowest value of the charts.

Tell a story or at least answer a question

Most initial data visualizers only create single charts such as graphs or bars. Sometimes a combination of two schemes is also useful. Let’s see how this is done.

You can find the data on Kaggle and Notebook. So here I made charts to analyze average sales of products, stores, and clusters. You can make different charts like pie and bar but I combined them all into one to get an overview of the analysis. Here we can clearly say that Store A has the highest sales and the most frequently purchased products are groceries and beverages.

If you think adding additional text will help the reader better understand the outline, just add the text. Let’s see with a real example. You can find this chart on Kaggle

The graph tells us which TV shows or movies have the highest ratings on Netflix. Here I have added some scripts like 97% of the audience like movies instead of TV shows. So when the audience reads the chart, they will know that the audience prefer movies rather than TV shows and they can compare ratings between shows.

Highlight some features such as changing the color of the higher value bar. Here while we’re talking about Netflix, I chose red and black for the chart instead of the simple white.

Working with a pie chart

I’ve seen many people use pie charts the wrong way.

Points to remember while working with a pie chart

• Never have more than 5 values ​​in a pie chart
• Always give a proper label, no matter how well you represent the chart, it won’t matter. Tagging directly on the graph is very useful as the audience does not have to search for myths. Finding legends takes time and we don’t want our audience to waste time on this.

In the example we see the percentage of shows watched on Netflix. We can clearly see that movies are a favorite here.

Choose the color of the pallet

NS boycott data, a qualitative The color palette works best with the screen. Custom colors should be easily marked to ensure accessibility.

NS Digital data, a sequential The color palette works best with the screen. Because numeric data must be placed in a certain order (ascending, descending).

a forked A color palette is a combination of two successive palettes that have a central value in the middle, usually zero.

The image below is taken as a color reference from Plotly.

end notes

We saw some common chart errors and how to overcome them with some examples. If you have any queries, you can contact me on any of these media.

Data visualization is an art form that needs to be mastered over time. Although these data visualization tips and techniques are not exhaustive, they will definitely help you move forward on the right track. Always remember that we make charts or visuals, not for our own understanding, we make these charts so that the audience can understand without going into technical matters. Understanding the audience’s perspective is the key to creating successful and effective visuals.

It doesn’t matter which tool you used to create elegant and well-behaved charts, it is important that we provide the core of the visual elements.

LinkedIn | Kaggle | average | Vidhya . Analytics

image source

1. Picture 1: https://www.kaggle.com/kashisrastogi/store-sales-forecasting
2. Picture 2 – https://plotly.com/python/builtin-colorscales/