IMPROVING VISUALIZATIONS FOR ACCURATE DATA IN POWER BI
This article is a component of my Power BI project, which revolves around rectifying incorrect visualizations. Our primary goal is to gain a thorough understanding of the data and implement appropriate visualizations to improve accuracy and effectiveness.
Introduction
In today’s fast-paced and data-driven business environment, it is essential for companies to have access to accurate and up-to-date information about their operations. One effective way to achieve this is by creating dashboards that enable managers to swiftly and effortlessly visualize key performance indicators and trends. In this project, we were provided with inaccurate visualizations for product sales, sales analysis, and customer segmentation, which appear misleading and unimpressive. We will be creating new dashboards using the same data and rectifying these incorrect visualizations to improve the clarity and comprehensibility of the data through visual representation.
1. Customer segmentation dashboard
Figure 1: Customer segmentation dashboard
A customer segmentation dashboard is a visualization tool that is used to display key metrics and data related to customer segments. Customer segmentation is the process of dividing a customer base into smaller groups based on common characteristics, such as demographics, purchasing habits, or geographic location. The main function of a customer segmentation dashboard is to provide a comprehensive view of the customer base, broken down by customer segment. This can help businesses to understand the specific needs and preferences of different customer segments, and to tailor their marketing, sales, and customer service efforts accordingly.
A customer segmentation dashboard can be used to track and analyze customer data over time, identify trends or patterns in customer behavior, and identify areas for improvement in the customer experience. It can also be used to track the performance of different customer segments or to identify opportunities for upselling or cross- selling to existing customers.
2. Sales analysis dashboard
Figure 2: Sales analysis dashboard
The function of a sales analysis dashboard is to provide a comprehensive overview of the performance of a company’s sales activities. It typically includes metrics such as sales revenue, sales volume, average transaction value, and conversion rate, as well as key drivers of sales such as product or customer segments. A sales analysis dashboard can help companies understand how their sales are performing, identify trends and patterns, and make informed decisions about how to optimize their sales efforts. In this dashboard, the value of discount, sales, profit, order quantity and total cost are placed at the top of the dashboard as a card which generally used as an overview for specific value and parameter. At the bottom, we can observe sales, by different customer segment, discount by product category and profit by product name and city.
3. Product sales dashboard
Figure 3: Product sales dashboard
A product sales dashboard is used to reflect product sales and profitability for the past period. Displaying information, such as trends and key indicators, enables managers to understand real product sales information to guide decision-making and achieve the purpose of data-driven decision-making. Sales and profit totals in this dashboard provide an overall view to give viewers a general perception of sales. The pie chart of profit by product name and product category shows the profit share of each product. Based on the profit pie chart, management can understand which products are the most profitable and which are barely profitable and thus adjust the product mix. Similarly, the treemap of profit by city and product name shows profits by city and product dimension, allowing us to understand the sales and profits of various products between different cities. Aggregating by region gives management insight into product preferences across cities, how sales strategies are being executed across cities and a host of other issues. At last, the stacked bar chart of quantity ordered new by month and product name represents the order volume of each product summarized by month.
This chart reflects the order trend from January to June and can be broken down for each product. In summary, the product sales dashboard reflects the basic situation of product sales and provides the data basis for sales decisions, i.e. data-driven decisions.
Issues
1. Wrong Chart Implementation
Using the correct type of chart is important because it helps to clearly and effectively communicate the data and insights being presented. Using the wrong type of chart can lead to confusion or misinterpretation of the data, which can lead to incorrect decisions or actions being taken. It is important to choose the appropriate chart type based on the nature of the data and the message being conveyed, in order to effectively communicate the information to the viewer.
In customer segmentation dashboard for instance, waterfall chart as shown in Figure 4 . Is used to show how an initial value is affected by a series of intermediate positive or negative values, resulting in a final value. This type of chart is often used to show the breakdown of a whole into its component parts, or to show the contributions of different factors to an overall change. In the context of a customer segmentation dashboard, a waterfall chart would not be an appropriate visualization because the focus of the dashboard is on understanding and analyzing the characteristics and performance of different customer segments, rather than on showing the breakdown of a whole into its component parts or the contributions of different factors to an overall change. A more suitable chart type for this purpose might be a bar chart or a pie chart, depending on the specific data and insights being presented.
This case also happened in product sales dashboard, using a Treemap for a product sales dashboard may not be the most effective visualization choice. While Treemaps can effectively represent the relative sizes of different categories, they can be difficult to compare and may not accurately display smaller categories. Figure on the left illustrate Treemap from product sales dashboard. It may be more effective to use a different visualization method, such as a bar chart or pie chart, to effectively communicate the sales data for each product
2. Unnecessary Volume of Data
Too much data is displayed in charts. Showing too much data is a problem that many producers often face. This problem is because the producer does not understand the problem from the user’s point of view but tries to show as much information as possible on the dashboard. This manner will cause viewers to get lost instead of giving them more helpful information. Take the pie chart in product sales dashboard as an example which is shown in Figure 6 below, it shows so many categories that it is impossible to see how many categories there are and the percentage of each category. The same is true for the other two charts in this dashboard
3. Indistinguishable Colour Scheme
The use of color scheme in a dashboard is important because it helps the user to quickly and easily understand the data being presented. By using a consistent color scheme throughout the dashboard, the user is able to easily identify different data points and trends. Additionally, using contrasting colors can help to highlight important data points and draw the user’s attention to specific areas of the dashboard. It is also important to use colors that are easy on the eyes and do not strain the user’s vision, as this can make it difficult for the user to effectively interpret the data.
As can be seen from the product sales dashboard, as demonstrated in right Figure. The stacked bar chart and treemap use colour to distinguish different products. However, it can be seen that some of the products adjacent to each other in the figure use colours that are similar. Such a colour scheme makes it impossible to understand data well in charts and even causes visual errors.
Enhanced Dashboards
1. Customer Segmentation Dashboard
Figure 4: Revised customer segmentation dashboard
In hindsight of the customer segmentation dashboard’s objective, it is essential to ensure that the dashboard effectively achieves this goal. To do so, it must provide answers to general questions, such as assessing the performance of each customer segment and determining the current value of these segments. Once we have these questions addressed, we can gain insights into specific customer segments and develop optimized marketing and sales strategies. Figure 8 depicts the new customer segmentation dashboard, and below are descriptions of some of the visuals within it.
Cards Cards are essential for displaying and monitoring numbers, and they are placed on the upper left part of the dashboard to provide an overview of key statistics. These cards include total sales and total profit, offering quick access to specific parameters. The numbers on these cards can be adjusted according to the selected filters, offering flexibility in data presentation. An example of this flexibility is demonstrated by the use of slicers for states and cities, positioned just below the card items. Slicers allow users to filter data based on their parameters, affecting the entire dashboard.
Scatter Chart The scatter plot showing profit and order value by customer segment helps us analyze how different customer segments perform. It reveals which segments generate higher profits and order values and highlights less profitable ones. This information guides marketing strategies and operational improvements. Additionally, the plot can uncover correlations and trends between profit and order value, aiding in identifying growth and optimization opportunities.
Table The table in the dashboard displays city data along with sales, order count, and order value metrics. It serves the purpose of presenting a clear and concise overview of this information, enabling users to compare and analyze different city data points. The table can help identify trends and patterns, and when combined with charts or graphs, it provides additional context and insights. Analyzing this table allows users to gain a deeper understanding of city performance, aiding in informed decisions about optimizing sales and operations in various locations.
Slicer The slicer by State and City, is a useful visual element had to be in our customer segmentation dashboard, it acts as a canvas visual filter where it enables a user to sort and filter with a city of interest. To see the customer performance or analysis in specific province.
2. Sales Analysis Dashboard
Figure 5: Revised Sales Analysis dashboard
Referring to the design presented in Figure 9, it is a sales analysis dashboard that offers a wide range of information. This interface provides comprehensive sales analysis, including data on sales, profit, quantity, and product categories over a specific period.
Furthermore, it offers insights into the company’s products, customer types, and location, including regions. The dashboard presents sales analytics for various customer segments, such as corporate, small business, home office, and consumer customers. It also displays the current discount percentages based on product categories like furniture, office supplies, and technology. Additionally, it showcases the volume of products sold and the corresponding revenue for each location within the active area.
To enhance data analysis, a visual slicer display is employed, allowing decision-makers to efficiently analyze data by order month and date. In total, seven different types of elements have been incorporated to improve the functionality of our dashboard.
Card We created three different cars that seek to show a number, such as the amount of sales, profit and discounts given. A Power BI card is a sort of visualization that is excellent for displaying such figures. Type of visual cards is used in card visualization, a participative technique for gathering data that enables groups to exchange and brainstorm ideas.
Slicer When we want to observe the dashboard’s overall visual display, the slicer is a good option. For quicker access, place frequently used or significant filters on the report canvas. Make it simpler to view the current condition of the filters without opening another list. Use the data table’s hidden and unused columns as filters. To provide more thorough date filtering in this modification, we used two types of portable slicers. The first one is by order date in days within the given period in the dataset between 1/1/2015 to 30/6/2015, The second one is by Weekend and Weekday, we came to this after creating two columns, one for Day of Week for each order date by using the WEEKDAY function built in Power BI, and the second column is Weekday/Weekend by using the IF Condition function. We should utilize the slicer tool since our goal is to have a stronger impact on the selective filtering of the data in the visualization during the period.
Clustered bar chart Next, we employ a clustered bar chart, which shows values or measurements with bars that are proportional to the data. Product category data are shown in this clustered bar chart broken down by city. Bar charts with clusters are useful for graphically illustrating (visualizing) our data. In addition to statistical indications, it is utilized. Multiple data series are shown in clustered horizontal columns in clustered bar charts. The horizontal bars are organized by city because each data series has the same axis name. Multiple series are directly compared within a particular category using clustered bars. The beginning of the chart is Washington have the high profit by city and product category.
Pie Chart The pie chart type is the last but certainly not least of the charts in our sales analysis dashboard. All versions of Power BI provide built-in chart visualizations called pie charts. Depending on the value of each data label, each set of categorical data is displayed in a pie form in a circular pie chart. Pie charts can be used to indicate percentages at specific moments in time as well as to display general percentages. Pie charts do not depict changes over time, in contrast to bar graphs and line graphs. a display of information regarding discounts based on product categories, for instance. As much as 55.14% of the second office supply category is visible, followed by 24.02% of the technology category and 20.84% of the kind of furniture product category
3. Product Sales Dashboard
Figure 6: Revised Product Sales dashboard
First, we adjusted the spacing between components in the new dashboard. As we can see from the summary data area in Figure 10, all components are bordered, and the negative space is unified, making the overall layout more aesthetically pleasing without being crowded or sparse.
Line and Stacked Column Chart Line and stacked column chart can show the number of quantity ordered of the products name by months from January to June. By looking at this visualization, it’s very easy to notice the interesting products by the customers and see the trending period of the products.
Clustered Bar Chart Next, we replaced the Treemap with a Bar Chart in the new dashboard because it is easier to make comparison. Treemap tastes confusing because it is difficult to extract useful information from it. The Bar Chart which shows profits with bars of product names and their categories. The horizontal bars are organized by product names as shown in this clustered bar chart listed as a descending manner by the profit, and broken down by product categories with different colors described as a legend. In addition to statistical indications, the quantities ordered and the remaining cards show more specific information about these products separately
Pie chart Lastly, we used a pie chart to shows the sales and profits of different product categories type, with the legend as product category and the profit as values, while we added the sales data as a tooltip. In response to the visual hierarchy, the style of summary data area has been changed to highlight the importance so that the viewer’s attention will fall on this area during the initial viewing. Guides viewers to explore the dashboard in order from overall to partial.
The overall layout of the dashboard has been adjusted, resulting in a more aesthetically pleasing instrument panel and a more balanced placement of components. People have a special obsession with symmetry, and asymmetry most of the time means unattractive. We Filled the background colour for the title area to make the title stand out and changed the colour of the title text to a more eye-catching white. Viewers’ attention is first drawn to the title when viewing the chart so that they are informed of the intended theme of the chart before viewing the data.
If you require access to the dataset used in this article or the .Pbix file, please feel free to reach out to me. I’d be delighted to assist you.