Introduction to Apache Superset

This introduction has the following learning objectives:

Editing this worksheet takes about two hours, depending on your previous knowledge.

No programming knowledge is required for this introduction; Basic knowledge of spreadsheets is sufficient for the time being.

For the exercises in this worksheet you need access to an Apache Superset service (see below), as well as a common web browser and an internet connection.

This guide refers to release 0.34 of Apache Superset, which was released in September 2019.

Good visualization is important for the success of decision making. Therefore, the findings of a concern, project or survey must be visualized. If findings are presented in a clear and comprehensible way, this increases their comprehensibility and acceptance.

Visualization can not only be used for illustration, but also for data analysis. Frequently, connections in the data can only be identified by a skilful representation. We humans are bad at comparing numbers; graphic patterns, on the other hand, we recognize well. Visualization therefore not only represents the data graphically, but can also be used as a separate technique for data analysis. The Internet also makes it possible to publish the visualization and thus to communicate easily with customers and colleagues.

Apache Superset is such a data visualization and publication tool. Some also call this application "Business Intelligence Tool". Superset is also a tool for sharing data sources, i.e. from table data to geodata. It can be connected to various databases.

Apache Superset is a web application and an account is required for access. If you already have access to Apache Superset, log in there.

Apache Superset knows several user roles that have certain rights to change data or call functions. In this worksheet it is assumed that you have obtained user rights corresponding to the "Gamma" role, i.e. you can create charts and dashboards.

The Geometa Lab HSR operates, among other things, "Apache Superset Cloud" as a "Software-as-a-Service". Apache Superset Cloud uses the PostgreSQL database system with the PostGIS extension and is subject to a fee (participants of Geometa Lab HSR continuing education courses receive their own access).

First you have to choose a data source. Under you can see the different tables. Here you can click on the magnifying glass to get more detailed information about the table,

such as which columns the table has and which data types can be stored there. In addition, you can set under Columns whether you can group by this column, whether it is temporal or filterable.

To query a table, click on the name of the table. In this example the wb_health_population is used. You can only select one table as source. If you want to link several tables to one data source, you need SQL knowledge (see other worksheets).

Selecting a table will open a new window respectively a new browser tab where you can create a chart. There you will find the selected table wb_health_population as Datasource.

Directly below, in Visualization Type (Chart-Type) you will find Table, which is a simple bar chart.

First you have to adjust the time span under Time, because the data of the table wb_health_population starts in 1960 and ends in 2014. We therefore use a custom filter of 2014-01-01 - 2014-12-31.

Under GROUP BY you can now define your first query by selecting a column like SP_POP_TOTL (Population Total) at Metrics and taking the sum of the columns. The Metrics column must always be used. It is also used on all representations to determine how much weight this row has (for a circle chart this would be the size of the section). If you now press Run Query, the query will be executed. It shows on one line the world population, i.e. how many people lived in 2014.

Under GROUP BY → Group by you can further subdivide the query. Group by is used to get from a total to an amount per selected attribute, e.g. the population per country. Therefore select the attribute country_name here (the preceding "ABC" shows the data type of this attribute, which is not important at this point).

Now save this chart as your first result in Apache Superset, e.g. with the name "Chart 1". You will need this chart again later.

Task 2
Now that you see the population of each country, we can limit the data to all countries with more than 100 million inhabitants. Try to do this using the filter under Query (you don’t need the _Custom SQL option for this)._

To make the whole thing even more beautiful, you can choose a suitable chart for the data under Datasource & Chart Type → Visualization Type.

Task 3
Take the result of task 2 and present it as a sunburst chart (Hierarchy is similar to Group by in Table View).

Now add region at the top of Hierarchy. The order is important for the display, so that you can see in which region the country and how much of its population it makes up. At this point it is advisable to remove or adjust the previously created filter so that you have more examples where you can see the consequences of the next step.

Next, we’ll adjust the color, because the graphic isn’t very pretty at the moment. The color of this chart can be adjusted in two ways:

Now save your chart under a suitable name. You will need it again later.

Next, it would be interesting to see population growth.

You can see the growth in Superset with the Big Number with Trendline Chart as a trend line, whereby this also shows the last value above the trend line. The chart would already be usable, but you can also set Options → Comparison Period Lag and Comparison suffix.

Comparison Period Lag_ specifies which years are compared to show the percentage growth. (If you type a five in Comparison Period Lag, it compares the last year he has (in this case 2014) with the one five years before it (2009)).

At Comparison suffix you can enter a text describing the percentage. The descriptive text is displayed directly after the percentage.

Task 4
Represents the population growth from 2000 to 2014 using the Big Number with Trendline.

Now save your chart under a suitable name. You will need it again later.

Since Chart No. 3 only shows us the average growth, it would be interesting to see the population growth per country to get a better insight into the distribution.

Task 5
Now, since you have already depicted the growth of the population, it should not be difficult for you to visualize the growth rate per country using a line chart. Set a series limit to limit the number of countries displayed and increase the time span to 1960-2014. You can type in the time span or set the filter to No filter.

Save this chart as well.

Task 6
Another way to display this data would be with an Area Chart. Try it out!

Chart no. 4 has already split the data very heavily, let’s adjust the chart so that the split is less strong to get a better overview.

Before you also save this chart again, let it show you the growth of the individual regions. In the attachment you will find a picture (figure no. 10) of the chart. However, the colors may differ.

Next, we want to try something more difficult, namely create a 6th chart as follows:

In this chart you can see how many inhabitants live in rural areas (X-axis), how high the life expectancy is (Y-axis), in which region a country is located (color of the bubbles) and how high the population is (size of the bubble).

This chart illustrates very well if and how life expectancy is related to the percentage of rural living.

To get this chart, first select the Bubble Chart and set the time span from 2011-01-01 to 2011-12-31. Then you have to select region under Series to color the bubbles accordingly. At Entity you have to choose country_name to show the countries.

Task 7
Select the correct columns for X Axis, Y Axis and Bubble size.

Also save this chart under a suitable name.

To finish, we will create this Chart No. 7, a Map Chart:

This map chart shows how many people live in rural areas by coloring the countries (black = 100% / white = 0%). In addition, a bubble indicates how many people there are. To see the percentage of the rural population or the size of the population, you have to hover with the mouse over the country or the bubble.

It’s a World Map. Here are the corresponding options:

Remember to also save this chart under an appropriate name.