1. Dashboards
Hi, I'm Kaelen Medeiros, and I'm the Product Data Scientist at DataCamp. Previously, you learned how Data Engineers collect and store data in a database. Now, you'll learn how Data Analysts visualize and explore that data using dashboards.
2. What is a dashboard?
A dashboard is a set of metrics, usually in the form of graphs, that update on a schedule. Some dashboards can update in real-time, but others update daily or weekly.
Let's discuss some common dashboard elements.
3. Tracking a value over time
One of the most common dashboard elements is a time series that tracks a value over time. This type of plot shows both the current value of a quantity and enough historical data to give context to that value.
In this example, we see the number of monthly active users over a ten month period.
4. Tracking composition over time
This new chart gives additional context to the previous monthly active users chart: it tells us the percent of those users that come from different sources, such as paid ads or social media.
The total percentages will always add up to 100 percent; that's why all of the bars are the same height.
With stacked bars, we can easily notice trends, such as the increase in traffic from Blogs in March and April.
5. Categorical comparison
Another common dashboard element is a categorical comparison using a bar chart. Here, we see the average number of minutes spent on a web page by different groups.
Whereas time series provides a historical basis for comparison, bar charts compare different groups during the same time period.
When you see a chart like this, it's always important to ask what data is included. In this case, we'll add a label indicating that this data is from the past 30 days.
6. Highlighting a single number
The previous three charts gave us many points of comparison for each value.
Sometimes we want to highlight just one number. In this case, we want to know how many people visited our website today.
Even when showing just one number, it can be helpful to give a small amount of context for comparison, such as how today's value differs from yesterday's.
7. Displaying text
Generally, we avoid adding tables to dashboards because graphics are easier to read.
Display text is an exception. For example, displaying a small number of customer comments can be a great way to add some qualitative data to an otherwise quantitative picture.
8. Where can we build a dashboard?
There are many ways of building dashboards. Some dashboards are built with a spreadsheet tool like Excel or Google Sheets. Others are built using specialized Business Intelligence or BI Tools, such as Tableau, Power BI, or Looker. For something really customized, some analysts use programming languages like R Shiny or d3 dot js.
All of these tools are great ways to get fast, accurate dashboards. It's best practice to ensure that everyone in the organization uses the same one so that there is no confusion about where to go for dashboard information.
9. Requesting a dashboard
Before issuing a request, be sure that a new dashboard is actually the best solution to your problem. Dashboards are required when you will need to access the information many times, that information needs to be updated frequently, and the information that you need will always be the same.
Once you're sure you need a dashboard, make your request as specific as possible. Do you need a single number of comparison over time? What time frame is relevant to you? Be sure to include your use case; this can help a data analyst choose the type of dashboard that is best for you.
10. Let's practice!
Now that you know the elements of a dashboard and how to request a new dashboard, let's practice!