Compelling Data Storytelling with Power BI: Titanic Survivors

This is a creative walkthrough of the Titanic Passenger Diagnostic Report published to the Power BI Data Stories Gallery. We'll walk you through the report and creative decisions that we made. Our data comes from the open source Kaggle Titanic Machine Learning from Disaster dataset. The report design is taken from the idea of an interactive museum exhibit, walking you through the night of April 15th, 1912.

On the left, we have introduced navigation as a way of progressing through the data story, taking you through an introduction, the demographics of the people that travelled on the Titanic, who survived and the survival rate. Adding navigation is a good way of identifying different topics within one subject.

It's a way of breaking in separate topics that build into one data story. If you would like to find out how we build navigation in Power BI, check out the link, How to Build Navigation in Power BI. in the description of this video. The creative approach that we have taken for this report is to introduce a sepia tone to the background to give it the look and feel of an old newspaper from 1912.

We have opted for sepia tones and strong imagery to set a strong contrast between background and the content. The cover page identifying the first branded screen. We designed the cover page like we do for all our reporting and to give a sense of endjin branding, so that we know where we are and in the right report.

It is always important to design a cover page as it's the first branded screen you see and sets the tone of the report. Make sure your company's branding is on the cover page. It doesn't have to be big, just a well balanced logo against the other content of the page design. Endjin. The introduction screen gives a brief outline of the report.

We then have an introduction outlining what the report shows, and how the disaster changed the future of maritime safety. We outline the report structure, and highlight how there were not enough lifeboats to save all on board. We talk about the role of the ship Carpathia, which arrived to rescue the survivors.

The demographics screen visualises all those that boarded the Titanic. On this page we are looking at where the passengers boarded and in what class 1st, 2nd and 3rd. In the top left corner you can see how many passengers boarded in each class. To the right we can see what proportion of passengers were in each class and the areas they occupied.

If you click the ticket class you can see which areas they occupied. The embarked location shows where the passengers boarded the Titanic. Cherbourg, France, Southampton, UK, or Queenstown in Ireland. The age shows you how many passengers boarded in each class with a passengers count and proportion male to female passengers in each class.

Creatively, we have utilized copyright free imagery, adjusting them in Photoshop for a stronger appearance. We have highlighted each chart by adding a subtle drop shadow, and taking the background tone down 80 percent to help lift the charts off the page and show some of the background imagery. The Who Survived screen shows who survived and who perished by class and embarkation.

On the Who Survived screen we are using a slightly different colour palette. We have based our palette off the red in the white star logo. If you would like to find out how to develop a colour safe palette for Power BI check out the link in the description of this video. Here we can see the who survived and who perished.

The perished count is marked in red. whilst the who survived is marked in blue. From the data we can see that there was not enough lifeboat capacity to save every one of the people who lost their lives that night, even with collapsible lifeboats. The survival rate shows the proportion of whether they were adults, children, seniors, teenagers or young adults who were saved by the lifeboats.

The plan view shows what proportion were male or female and what travel class they were from with a stark reminder that there was a 38 percent survival rate from the disaster. This concludes the creative walkthrough of the Titanic report. This Power BI report was shown at SQLBits 2024 when Barry Smart demoed the report as part of his Microsoft Fabric Machine Learning tutorial.

Which he has turned into a series on this YouTube channel. When designing your report, you need to think about the medium it will be presented on. As this report was being debuted at the SQLbits conference, what was designed on a monitor was presented on a projector. We opted for strong contrast between foreground and background artwork.

Sometimes, what may seem strong on a monitor can be lost if there is contrast on the projector, especially when you factor poor and harsh conference lighting which can adversely affect the contrast.

We can zoom out on this to get a wider perspective of its location. Below this, we track the population growth or decline shown as an area chart with a solid line. The lines we use are as important as the color palette we have develop. As each line style represents the same metric as we go through the report, we go back to the overview page and look at another topic.

In the health section, we have utilized a series of area charts looking at the median health color coded by region, showing life expectancy from birth in years. When we look at the distribution of health, you'll notice we have top and tailed the min-max life expectancy with our dotted line. From the previous section, adding muted dash lines to the metrics in between, while it's also utilizing different marker shapes for better legibility.

On the health gap chart, we have applied the same system to differentiate the content. If we go back to our overview page, we can see how this design approach works in the wealth section. We have changed the chart slightly to use line charts, looking at our median wealth color coded by region, the distribution of wealth by decade with our dashed and dotted lines, and we can see the wealth gap between the richest and poorest countries.

If we return to our overview page, there are a couple of sections we'll let you explore. However, we have covered the design aspects that we set out at the beginning from branding, layout, interactivity, data stories, and our exploration into developing an accessible color palette accompanied with other methods of visualizing.

An important element for consideration when we develop our reports is not overloading the pages, so we look for clean, clear layouts to allow our content to breathe whilst conforming to the foundations of our grid.

If you have any comments, please get in touch. All feedback is welcome, and thank you for watching.