 In my previous videos, I have used human development index data to illustrate some of the basic mechanics of Orange, widgets and workflows. I will also stay with this dataset in this video and conclude our overview on visual programming in Orange with a couple more widgets for data visualization. We again load them with the datasets widget. Orange remembers my previous use of this widget and loads HDI datasets automatically. Remember, HDI includes data on 188 countries. I am interested in the distribution of values for life expectancy. Let me use the distributions widget. I will choose life expectancy from a list of variables and decrease the bin width to increase the resolution in the bar chart. The distribution is skewed to the right. There are more countries where people live longer than countries where the expected lifespan is very short. Let us see which country has the lowest life expectancy by clicking on the corresponding bar. The distributions widget emits the selected data items. If I use the data table, I can see that life expectancy is lowest in Swaziland. Most European countries and several countries in East Asia are where people live longest. There is another widget that displays the distribution of feature values. It is called a box plot. Here I use it to examine some statistical features of the distribution of life expectancy values. Our datasets mean life expectancy across 188 countries is 71.3 years. One half of the countries have a life expectancy from 65 to 77 years, the numbers denoting the first and third quartile of the life expectancy distribution. To learn more about this particular widget and visualization, I can always bring up orange help. Here, for the box plot, I can learn that the box plot also displays the median and standard deviation. Now, here is my plan. I want to use the distributions widget to select the countries where people live long, say above 70 years. And then I would like to use box plot to find the differences between the selected longevity countries and countries with short life expectancy. For a start, I use the distributions widget to select 118 countries with a life expectancy of over 70 years. Notice that there is an info text in the status bar of this and all other widgets. Clicking on it shows me some core statistics of the outputs of this widget. The distributions widget has three output channels and the one I am interested in is the second one, which includes all the countries and the feature that reports if the country was selected. Let me inspect the data on this channel in the data table. I will double click on the link between distributions and the existing data table and tell orange to pass the entire data set instead of passing only the selected data. The data table now includes a new column, selected, where the value of this feature is yes if I have selected a country in the distributions and no otherwise. I will use this selected feature to compare the distribution statistics of the two groups of countries. Let me connect a new instance of a box plot to the distributions widget and rewire the connection to pass it to all the data. In the box plot, I chose selected as my indicator for the subgroups. Notice how the statistics are now reported for the longevity countries marked with yes as they were selected in the distributions widget and the other countries marked with no. I am looking for the features where the difference between the two distributions as measured with students t statistics is the largest. Box plot in orange has a neat feature where I can order the variables according to the relevance to subgroups. I will switch this feature on and features related to life expectancy make it to the top. This was of course expected. We also expect that mortality rates would be related to longevity, but we also see that longevity is correlated with mean years of schooling and human inequality. In countries with lower inequality, people live longer. I have introduced two new widgets, the distributions and the box plot in this video. And we have again learned about using different kinds of widget output and about using box plots to explain the differences between two groups of data instances. I will also use box plots variables ranking according to the subgroup indicator in my following videos when we deal with clustering. That's right, clustering is coming next.