The DataViz Process, Part 2: Exploring the Met with Flourish
Are you interested in learning more about how to create data visualizations but not quite sure you’re ready to step into the world of Tableau, the software platform that is becoming an industry standard in the world of data analysis?
Why not try Flourish?
The browser-based application allows you to create vibrant and engaging data visualizations in a large variety of formats and styles. It also lets you add some basic animation to your visualization, turning your presentation of a static dataset into one that really helps you to capture and demonstrate the way role of time in the trends that you’ve uncovered. And, like Tableau, Flourish has a story mode where you can put together a dynamic presentation incorporating multiple visualizations to really help show the scope of your data and analysis.
To get started with Flourish, you just need a free account and a clean dataset (see Part 1 on cleaning your datasets). Flourish does offer subscription accounts that allow you to keep your visualizations private and publish them without a watermark, but for most things you might need to create a graphic data presentation for in college, the free account should suffice. Once you sign up, you’ll be able to get started designing immediately!
The Dashboard
When you first log in, you’ll see your Flourish dashboard. Here you can start a new visualization, start a new story, as well as edit and manage your existing projects.
While some freemium applications limit the number of projects you can have at any one time, Flourish doesn’t put any limitations on a free account in that way. So you can create as visualizations and stories as you like without worry.
Creating a Visualization
I’ve had experience using both Microsoft Power BI and Tableau, and while they’re great programs, they can be a bit intimidating to new users. But I’ve found Flourish to be very user-friendly, even for someone who has little to no experience working with datasets and visualizations.
One of the reasons I find a great alternative to Power BI and Tableau is their visual presentation of the different templates that users can choose from. Everything from the kind of line, bar, and pie charts people are most familiar with to 3D maps, pictograms, and a number of different kinds of diagrams. And you can hover over a template with your mouse to see a little blurb what kind of analysis a template is used for, or use the ? icons for a little more information about how to use them.
In Flourish, you begin by choosing a template. This is different in Tableau, where you begin by uploading your data, and while Power BI will let you select a visualization style first, it doesn’t offer any hints on what kind of analyses different kinds of visualizations are best suited for. So unless you’re well-versed in different types of graphs, you might not know the best type of chart or graph to choose for your data.
I chose a bar chart template to start with. Once you’ve selected your template, you’ll see your project dashboard appear, pre-loaded with some sample data. You can upload your own data and replace the sample data using the blue Upload Data button (see screenshot below). Once you upload your data, you can start telling Flourish what you want to do with it.
This is where I think Flourish is great for people who are just getting started with data visualization. Sometimes when working with a program like Power BI or Tableau, it’s a little hard to understand what data (rows or columns) correlates with what field. But in Flourish, the color-coded fields and the clear and simple labels really help you to put together a professional-looking visualization quite easily. And the program only includes the fields that are necessary and important to the type of visualization that you selected.
If you look in the screenshot above, even though Flourish tried to anticipate what I would want from my dataset, I do have to make some changes. First, I want to change the Labels/time field to correlate with column B, the Accession Date data —
But, here is where I’ve hit some snags with Flourish.
- Size: even after I’ve cleaned it and gotten it sized down, the dataset is still pretty big. Flourish will upload my file, but then everything gets slowed down and I get a few “Page not responding, wait or exit” messages from my browser. This can be really frustrating.
- Messy Data: for what I want to do with it, is still kind of messy. It could be a lot tidier (another word for clean data).
So now I’ve got to make a choice. I can continue to wait and try to get Flourish to work with my dataset as it is now, or I can go back and try to clean my data a little more.
The Aggregating Step
One data cleaning step we didn’t talk about in the previous post is aggregating, or summarizing. And that’s because aggregation is helpful on more of a case-by-case basis.
My data is still raw. What that means is that even though we’ve taken some steps to process and clean it, it still exists in its original observation form. Remember, observations are another way to refer to the individual rows of data. Each individual row that I’ve retained after cleaning out the ones I didn’t want (anything that was added to the collection before 2000) is still in my dataset as an individual row of input. This is why my dataset keeps clogging up my browser’s speed.
What I want it to look like instead is a column of years, followed by a column for each department, because this is the variable being observed — which department each object is assigned to.
To do that, I’m going to use the PivotTable feature in Microsoft Excel (stay tuned for a post on how to use PivotTables!) to aggregate my data into a summary table like the one on the left.
Now when I upload my data to Flourish, there’s no lag. And I’m immediately able to start working on my data visualization.
Creating My Met Data Visualization
Now that my data is ready to go, I want to make sure that the right info is feeding into the right criteria. You can see on the right of the screen capture above that there is a dashboard where I can indicate which column should be interpreted for the labels, which for the values, etc.
My label/time information comes from Column A, the years 2000 through 2021. The values are going to be the number of items added to a collection each year, or Columns B-W. Each chart type will have some specific criteria that you can choose and select as well. But for what I want to visualize, I just need to indicate the labels and the values.
And what is really great about Flourish is that even when you’re working on the data side of your dashboard, you can always see a little popup of your visualization on the lower right, immediately reflecting any changes you’re making. But if you want to see your visualization in detail, you can use the “Preview” button at the top of the dashboard.
But once I look at the preview, I can see that my visualization looks really busy, and somehow manages to be both too compact and have too much white space at the same time. I don’t like all the labels at the top, and it’s really hard to see the details of how departments acquired new material over the last twenty years.
There are some things I can do, some questions I can ask, and some choices I can make, to make my visualization clearer and more useful.
I can play around with the template options on the right-hand side of the dashboard. There are numerous ways you can change how data is represented.
Some options include adding a filter or control, so that only one department’s data is visible at a time. Except what I really want to do is see a comparison of departments, and so using a filter won’t allow me to do that (you can have a “Multi select” filter to view a few categories at a time). But if you want people to be able to interact with your data visualization, or a filter fits the question you’re asking, that’s one way to make your viz look cleaner and more streamlined.
Another option I have is to consider my data again. Could I aggregate further? Perhaps group the years together in blocks of four or five years? Could I remove some departments that I’m not really interested in, like the Libraries or the Costume Institute or the Musical Instruments? Perhaps, but if I do that then part of my original research question, or my contextual discussion, is going to have to address why I narrowed down the departments.
But what if I’m not using the right template? A bar chart is just one of the many templates that Flourish offers, maybe a different template will better represent my data?
It’s easy to switch between basic graph or chart templates using the Chart type selector on the right side of the dashboard (expanded in screen capture below). And I recommend trying a few different ones until you get something that works for you. (Don’t forget to go to the Data dashboard, too, to see if there are special or unique requirements or considerations for different types of charts.
I tried a number of different template types until I found one that really highlighted the answer to the question I was trying to ask:
Which departments of the Met acquired the most new material, and when?
And the final chart style I settled on was the stacked bar chart with the values of my data expressed as a percentage of a whole (or the number of items added to Photographs in 2006 expressed as a percentage of all the items added to the Met Museum in 2006.
Fine-Tuning and Customizing
Once you’ve settled on the best form for your data visualization, Flourish allows you to customize many of its elements. For example, I decided against using a filter, again, because I felt the addition of it was unnecessary, especially if my chart was going to be a static image. And I liked the way the legend, though a little large, clearly expressed which color signified what department.
Other things you can customize include setting Max and Mins for the X and Y axis, the colors and color palette, how thick your bars (or lines) are, how transparent or opaque. You have some control over the order of the layout — title, legend, plot, etc.
If your data viz is going to be embedded or linked to online, you can add interactive elements. The filter control is one, but you can also add popups to your data, and specify what information should be included in them.
Some types of charts can be animated — either the chart itself (the Bar chart and Column chart race templates are a good place to start) or from slide to slide in a Story that uses the same dataset.
Publishing and Exporting Your Data Viz
Sharing your visualization is a snap. There’s a button at the top right of the Flourish webpage where you can Export & Publish your creation. Flourish will then create a unique link to your data viz that you can share out on social media or send via email. And it will also provide you with the embed code to share out your visualization on your personal website or other webpage. (It gives you embedding options, too: Script, AMP, or iFrame.) This option is great for visualizations that are interactive or animated.
Pro-tip: To embed your viz in Medium, you can just paste the link and Medium will automatically transform it into an iFrame embed of your chart.
But you can download your visualization as an HTML page or an image (PNG, JPEG, or SVG) too, though interactive elements will not be preserved in this format.
Once you’ve published your creation, you can continue editing and making changes to it, but you want to remember to go back to the Export & Publish button to republish your work with those changes. The link and embed code should stay the same, and your changes will then appear. You can also use this button to unpublish your visualization if you decide you don’t want it to be accessible any longer.
What Next? Stories
So now you know a few cool things that you can do with Flourish.
But you also know a lot more about the process of creating a data visualization, beyond just using Flourish. Including how to find a dataset, clean a dataset, what to do when your data isn’t working for you, settling on the correct way to present your data, and how to share it out.
In our next post, I want to share how to use Flourish to create an engaging data visualization presentation using their Stories feature.
Stay tuned!
