TycheAbout:DataStoryComponents
What are the components of a data story?
A data story has 5 components:
- An idea or a question.
- Government-sourced data relevant to the idea.
- Honest analysis, done with an open mind.
- Clear presentation of the analysis results.
- A point of view: conclusion, question, comparison, or confusion.
The idea or question, and the relevant data can happen in either order. Sometimes you have an idea and you have to go find relevant data (Tyche helps). Sometimes you have a data set and it inspires an idea. Either works.
Honest analysis and an open mind are essential. The honesty test is that finding a piece of data inconsistent with your idea means you’re happy to rethink the idea, not ignore the data. The open mind test is that you’re equally happy to write a story 180 degrees from what you thought it would be at the outset if the data takes you there.
Clear presentation means visualization (Tyche helps). Words are powerful. But they are not enough alone. Both the raw facts and the point of view need to be presented in a way that very busy readers and (potential) users will absorb quickly and correctly.
Data stories end with a point of view. Not an opinion. Not a pre-conceived notion. Not a talking point. A data-driven point of view fully supported by all the available facts. That point of view typically takes one or more of four forms: a conclusion, a question, a comparison, or confusion.
A conclusion is when the data is complete and clear enough to justify plain statements and deductive reasoning. You are going to make a statement based on what you learned. If you were honest and kept an open mind, it might not be the statement you thought you’d make when you started. An example is that a certain city’s contract with a vendor does not have the intended outcome.
A question is when the data either only or also logically leads to one or more questions to be asked next. Those might lead to subsequent data stories. They might be the call to action for someone to use your story (maybe you!). An example is that after completing and presenting analysis of the distribution of a certain tax incentive, you might ask “was it worth it?” and lead to a follow up story.
A comparison is when similar data sets have been analyzed for more than one reasonably comparable geography, enabling them to be juxtaposed. That in itself might lead to conclusions or questions. An example is whether similar cities with that same sort of contract or the same sort of tax incentive had better or worse outcomes.
And confusion is when you can’t reasonably get to one of the other three options. Confusion might lead you to sit on a story until it can be more complete. But sometimes confusion is actually the story to tell. An example is if the government data provided is so disordered, incomplete, or otherwise unusable, you can’t get to a conclusion, question, or comparison. Your point of view might be that the locality in question has some work to do on its open data policies. That would be better supported if you could make a comparison to how similarly sized localities do in the same regard. Maybe there’s an opportunity to enlist a broad community for action.
Next: What is Tyche Insights?