Over the past couple of months, the visual analytics team at Decisive Data has participated in two data visualization projects run by Viz for Social Good (a place where non-profits can get help from data visualization experts and where people that love data can practice their skills for a good cause!)
You can view the work we’ve done so far here:
Participating in Viz for Social Good has been an awesome experience because it’s an opportunity to
- Learn about great social enterprises all around the world and…
- Team up on a creative project that helps those enterprises.
We encourage other visual analytics teams to participate as well! You can get involved here.
As part of each project that we participate in, we’ve decided to also release a blog post with a few details on the process we take to create our submissions.
Here’s a look behind the curtain for our second Viz for Social Good contribution: a dashboard to promote awareness of homelessness in the UK for Tomorrow Today.
Viz for Social Good: Tomorrow Today
This Viz for Social Good project is for an organization called “Tomorrow Today”, organized by UK activist Papakow (“Papa”) Baiden.
Tomorrow Today seeks to bring awareness to and end homelessness in the UK.
See a quote from the founder below, and learn more here.
“Homelessness is not acceptable. Not in 2018, and not in one of the richest countries in the world; the UK. We can’t end it overnight. Nor can we do it on our own. But [Tomorrow Today] lays out our motivations, scale of ambition, and plan for how we will lead the change in a unique way as an agile, independent social enterprise making policy change happen, whilst creating the public interest to make the conditions for change more likely.”
Data Discovery and Whiteboarding
After reviewing the main goal of the visualization, to help people understand homelessness in the UK, we started to dig into the data provided to us.
There were two main datasets. One included data per local authority for the number of rough sleepers, and another included vacant housing information by local authority. Below is a quick outline of some key discussions/decisions we made leading up to our whiteboard.
- Pivot the years! The dataset originally had years as columns across the top. Having data structured like this makes it difficult to see values trend over time in Tableau. Therefore, we pivoted the data using Alteryx Designer.
- Identify the key measures. There were two main metrics in the rough sleeper count dataset: both rough sleeper count and rough sleeper rate per 1000 households. In our final visualization, we decided to display both.
- Map those local authorities. Hailing from the states, we weren’t quite sure how to map the local authorities. Luckily, we found some helpful resources here that allowed us to get this all in order.
- Do we use all the data? Datasets on both vacancy and rough sleeper rate/count were provided. We chose to focus on the rough sleeper rate/count as it was difficult to point out a concrete relationship between vacancy data and rough sleeper data.
- An up-close look at London. Because London is a population center and has very granular local authorities, we decided to pull it into a separate section where we could zoom in.
After much discussion, we created a whiteboard to outline our ideas and prepare to create our visualization in Tableau!
Viz in Tooltip
We found a great use case for Viz in Tooltip! In addition to showing the overall picture of rough sleeper rate per county (the map), we wanted to show detail for the local authorities within each county. Viz in Tooltip allowed us to provide that extra context without adding additional visuals to the dashboard itself. The audience can hover over any county and view the rough sleeper rate at the local authority level, as well as see the trend of rough sleeper count.
An “easily digestible” viz
There was a lot of information to present! Right at the beginning we had a discussion and made up our minds to keep our overall viz as simple as possible. We wanted to make the viz personal by showing a map of the UK, with annotations of interesting data points. At the same time, we wanted to keep it simple, which meant resisting the temptation to add more charts. We definitely had to decide what was most important to show, even though it meant discarding other good ideas we had.
Custom color palette
We liked the idea of blue for the map and orange for the contrasting visuals (rough sleeper rate vs. rough sleeper count), but we wanted to make sure the blue stood out and was engaging, even to users familiar with the Tableau default blue! After careful thought, we finally selected the unique blue palette in the final viz.
We used Color Brewer and Tableau’s custom color palette feature to create the unique color palette for our map. It's a great resource to find coordinating map colors.
Thanks for reading, and we hope you’re inspired to join Viz for Social Good as well!
We'd love to hear your thoughts—reach out to us on Twitter @DecisiveData and stay tuned for more socially-themed vizzes!
Posted by Alyssa Hudon, Chris Herron, Jacob Olsufka, Joshua Banks