1. Visualizations for different audiences
Great job on these exercises! Let's learn about how to include visualizations in our story.
2. Communication strategy
We know how to craft a story for a technical or non-technical audience.
3. Communication strategy
Now, we are ready to choose a data visualization intended for that specific audience.
4. Data storytelling
We have seen that visualizations are key elements for a data story.
When we find the insights and have the right data, we still have to choose and adjust a visualization to the message we want to convey. Again, we should consider our audience
expertise in the topic
and familiarity with the concepts, to select what graphs they can easily interpret.
For example, a density plot would be OK for a technical audience, but a histogram would be more appropriate for non-technical people.
5. Tailored message
Let's get back to our chocolate project to illustrate some best practices for visualizations.
Imagine we need to communicate two different messages:
the first one for an investor,
and the second one for our technical lead.
6. Directly linked to message
To include relevant visualizations, we should determine which data elements are required to help us convey our message, and only include those visualizations that are directly linked to our message.
For our investor, we could show a graph showing the forecast monthly revenue for both scenarios: launching and not launching a marketing campaign.
For our technical lead, we need to include a graph that shows how well our model performed on historical data.
7. Provide context
Visualizations should also provide context to support our message.
For the investor, we can include a graph showing the most influential factors for a customer to buy our chocolates. So for example, the higher the price the less likely the customer is to buy, and vice versa.
The technical lead will be interested in this, but also more details on the analysis itself, such as the detailed feature importance.
8. More best practices
The Pareto principle
states that the majority of outputs come from a minority of inputs. So outside of the data contributing to our story, there will be some data that is less relevant but that we want to include. We should aggregate it to reduce noise, such as
presenting the sales for chocolate, chips and other products instead of each product separately.
Our visuals can be made approachable and engaging by
considering our audience background, and simplifying the visuals to the audience knowledge level.
Basically, you care about how many insights your audience gets form your visualization, and how quickly they get it. A complex plot gives many insights but takes time to understand; a simple plot gives few insights but is quick to understand.
In general, less is more. Instead of showing a lot of detailed visuals, we should just focus on the most simple and relevant ones that support our message.
9. McCandless method
There are some steps that we can follow to include and present visualization effectively to our audience. They were established by David McCandless, a British data-journalist.
First, we should give our graph a headline, which we then use to introduce the visualization and focus the audience's attention on it.
The headline should be short, clear and obvious: it supports our story and clarifies the visualization. In doubt, we can use the y axis vs x axis technique, for example: "chocolate sales by month".
10. McCandless method
Then, we should anticipate common questions from our audience, such as "Where does this data come from?" or "Why did you focus on this characteristic specifically"?
With these questions answered before being even asked, attention is focused on our graph and our story.
11. McCandless method
After that, we should clarify what insights the visual is showing. We should explain to the audience what they are seeing, and not assume they will ask later or understand by themselves.
For example, chocolate sales have decreased with time.
12. McCandless method
Lastly, we need to help the audience relate to the graph and its insights.
making sure that they understand why the insight matters, how it will support further insights in the presentation, or how this insight can be acted upon.
For example, the chocolate sales are relevant to the finance department, as they impact the revenue, or to the marketing department, as advertising might help.
13. Let's practice!
Now it's time to practice these concepts.