There’s no point investing in big data if you can’t visualize the insights
For many small or newer organisations, the collection of data is task-based. Sales information is collected in order to invoice the customer and to update inventory. Staffing information is collected to calculate pay and meet human resource requirements. In the early stages of an organisation’s life, this task-based data collection is all that is needed to meet their needs. But as the organisation grows out of survival mode and wants can be considered along with needs, they can look to their internal resources to see the source of strength they have in the data they have collected, and that they can use those data to report on the activities of the organisation.
Then the questions become, how do we use this data? How do we question our own organisation? Are we functioning well? Are there areas where we can improve? How do we present these data to make those answers clear?
Not everyone has the ability to see or understand the information they are looking at by viewing lists or tables of numbers and words alone. And even the most mathematically inclined or data savvy person may not see certain patterns or trends existing in the data unless they can see a graphical representation of those data. Sometimes the best way to see what our data is telling us is by using data visualisation.
When we hear the term “data visualisation”, we may think of beautifully designed infographics that are more complex than what our little organisations need right now and requiring software to produce that our little organisations can not afford right now. But data visualisation does not have to be complicated or expensive. It is simply any method of presenting information in a way that can be interpreted visually.
You may be more familiar with data visualisation than you realise. An election map showing results color-coded by region is data visualisation, as is a tree diagram like a family tree, or even a subway map. Any time information is taken from a list or table and graphically presented in such a way that a person can see, at a glance, what the data is saying, that is data visualisation. These images are such a part of our day to day lives that we do not recognise them for the data they are rather than just pictures.
For the purposes of a beginner, the most common data visualisation tool already exists on most organisation’s computer systems – MS Excel. For those who do not have the Microsoft Office suite of programs, there are low-cost, or free, alternatives available online that have many of the same features as Excel. As this is a guide for beginners, most of the beginner functions would be available on these lower cost alternatives.
What’s Suitable?
The most common method of data visualisation is to organise your data in charts/graphs. The three most popular, and the three most likely to be useful for a smaller organisation are: Bar, Line, and Pie charts. Each of these have their own strengths and are best suited to particular purposes.
Bar charts are best for comparisons. If you would like to know the differences in the number of sales for each month, a bar chart would be the best choice. Along the vertical axis (y-axis) would be a scale showing the number of sales, and along the horizontal axis (x-axis) would be each month. In the chart area would be the bars of differing heights to reflect the number of sales. Bar charts can be vertical or horizontal. They can be grouped, for example to show more than one year, or stacked to show more than one product type.
Line charts are best for showing trends over time. Showing the number of sales over a long period of time – say five or more years, with sales on the y-axis and dates on the x-axis, a line chart can show whether the trend had been up or down without having one very bad or good year influencing the way the growth appears to be heading.
Pie charts are best for showing percentages of a whole. A table showing sales by region could show that the Northern region had 27% of the total sales and the pie chart would have a slice labeled “Northern” that was 27% the area of the circle in addition to the slices and percentages for the other regions.
No matter what type of chart you use, you must first do some data clean-up to ensure that there are no duplicate values, erroneous entries, or unusable rows in your table. Once you are sure that your data is clean and accurate, you should think about how you would like your chart to look, what data you would like presented, and plan your source table accordingly. Once you are satisfied that your table is clean and contains all the data you plan to report on, most newer spreadsheet programs will have an “Insert” menu option where you can find the different types of charts that you can insert into your spreadsheet.
Have a Back Up
It is at this stage that you can pick the type of chart you would like to use based on what type of reporting you wish to do – comparison, trend tracking, percentages, etc. You will also see options for other types of charts – scatterplots, sunburst, waterfall, etc. These are all useful charts, and ones you may find illuminating. If you wish to choose one of those types, just to see what they tell you about your data, do so. As long as you have a backup of your data before you begin playing with it, feel free to explore the different types of visualisations if you have the time. You may find a reporting method that works perfectly for your organisation that is not one that is commonly used by other organisations. Creating visual reports should not be intimidating. If the chart you choose is not one that works, scrap it and choose another type. Again, remember to backup your data, especially if your spreadsheet is being used as your database (the place where this data is being held, not just manipulated).
In newer versions of Excel, the program will even suggest chart types based on what kind of data you have in your table. You do not have to maintain the formatting that has been assigned to the chart, you can change colours, whether there is a border, font size, etc. A good rule of thumb with visual data representations is to keep it as simple as possible. It may be tempting to make the chart very visually interesting, but we want the focus to be on the information that is being presented, not on how the information is presented.
The ideal data visualisation is one that requires no further explanation than the image itself. If you maintain your data and use the correct charts to answer the questions that are asked, you should be able to do so quite easily. And as the depth and complexity of your data grows, so too can the options for presenting it, and your comfort in using those methods.
