Makeover Monday 2020 W2: Use of harmful pesticides in US agriculture

Makeover Monday Week 2 was an interesting exercise for me in a few points: the data was quite challenging, I had a play with risky infographic style design, I learned how to do bar charts with filled shapes, it got fun mixed feedback - the best kind of feedback!

This blog post is a brief exploration of these points while going through my process to create it.

The original viz:

The original visualisation comes from a research paper, published here, which aims to shed light on the use of pesticides by the US, that are considered harmful in other countries. It is quite a fascinating topic, and the paper is filled with tables and visualisations while trying to explain the data behind the authors' findings. The chart below is the first presented in the Results section of the article.

What works well:

• It is a simple bar chart, choosing to display just a top 3.

• The labelling inside the bars is helpful while you are reading the article and trying to relate it to the image.

What could be better:

• The chart mixes what is essentially two separate metrics in one chart, but there is no differentiation of colour or spacing.

• The gradient blue is a bit unnecessary, but it doesn't take away from the clarity of the chart.

• The labelling for the last 3 bars is very confusing. It took me a while to get what it actually meant, and even though it is within the context of the paper, it is still difficult to understand it at first sight.

• European Union - a group of countries - is compared to individual countries (China and Brazil). On a side note, I understand that those two countries are not comparable in size to European Union ones or the US. However, it still feels inaccurate to compare individual nations to an international political and economic group.

My approach:

1. Understanding the data & my potential audience:

This was a tough one. The data shared by the Makeover Monday team shows the quantity of pesticides used in the US in 2016 (lbs) and a percentage of those that were used in the European Union, China and Brazil. It is a simple enough table, but it didn't match the chart, which displays the number of pesticides. As time was a pressing matter for me, I decided to accept it and go with the readily available data.

While trying to come up with a story out of it, though, I noticed how that table actually gives me a minimal view of the issue treated in the paper and how much I wish I had time to fetch a more significant data set. I had so many questions: how many pesticides are banned in the US and not in other places? Does it mean that there is an overlap between the number of pesticides on each of the other regions? When they say all three regions in the last grouping, is it at least in those three or those three alone? The list goes on.

Then comes the question of who would be seeing my viz. The short answer is the general public - a non-homogeneous group of people with different interests that might or might not be concerned with the use of pesticides in their food. One could argue that the audience should be the readers of the paper, and that makes for an entirely different viz (more on that later).

A general undefined audience is a tough one to reach, even if the message is incredibly simple - the US is lagging behind other places in banning pesticides that scientific evidence and other regulatory agencies already consider harmful. In simpler terms: people in the US end up eating more poison with their food than outside of the US. Oh!

2. Defining which story I'd like my visualisation to tell:

Since I kind of understand that two separate things are going on in the original chart and the dataset provided, I decided to just go with the simplest one: the US used approximately 1.2b lbs of pesticides in 2016, of which 27% are banned in the European Union, 3% are banned in China, and 2% are banned in Brazil. That statement alone doesn't leave me much to work with. It is essentially one big number with three supporting details from it.

The idea that if a pesticide is banned in a different place, it is basically considered a hazardous substance for life (like poison) caught my attention. I could play with that and try to make the visuals more interesting than the lack of a more in-depth story to tell.

Important note on data visualisation best practice: I understand this is a polemic point of view. Design shouldn't overwrite meaningful information, nor it should mislead people. Ever. In this case, the data is very simple and straightforward. So simple that if stripped out of all flair, it is essentially one number with three additional pieces of detail near it. If I were required to design this for the researchers, accompanied by a paper detailing facts, I probably wouldn't go the route I did.

But this is Makeover Monday, and it is supposed to be a fun and quick exercise. I have tried infographic-style visualisations before, but I've never had the chance to go more experimental and expand my use of iconography in dashboards. I thought it would be a fun little exercise to try and pull off an icon-filled bar chart.

This is what I wanted my viz to say:

• The US used 1.2b lbs of Pesticides in 2016;

• Out of these, a portion is banned in the EU, in China or Brazil;

• If you are on a metric system like I am, this doesn't mean much, so I'll covert it to tonnes;

• If a pesticide is banned, it is considered to be hazardous for humans (like poison!), so I might explore that.

3. Getting creative & Sketching:

I sketched a lot on this one. I wanted it to be interesting, still simple to do, I wanted to try icons, but I did not want to end up with something that was just too much.

This is where I landed:

4. Building it in Tableau:

The first thing I did was to add an extra column to the data, converting lbs to tonnes. If I'm supposed to have a broader reach, this means I can't ignore the fact that fewer places use the Imperial system.

Next, I added a switch to flick between the systems. It is done using a simple string list parameter, as a single choice format:

Then, I tie the measure to the parameter through a formula, that will change, based on the choice selected. This is the measure used to coordinate the numbers in the viz:

Now, to the fun feature: the filled poison flasks!

The first time I saw this type of visualisation in Tableau was on this tutorial. It has been sitting on my bookmarks of cool Tableau stuff to try for a while, but I never got the opportunity to do so (I worked with Supply Chain and Finance previously, so let's just say creative visuals are frowned upon). Given the theme and straightforward story to tell from the data, it seemed like an excellent opportunity to give it a try.

The first step was to find an icon. I had a rough idea that I would need something that is vertical enough to resemble a bar - like a chemistry tube - but that would also resemble something hazardous. I usually use flaticon to search for icons, and this is the one I used:

I then used Photoshop to change the outline colour to white, and paint the outside area to match my chart's background. This is the trick: just leave the part you want to be "filled" as a transparent background. Every other part of the icon needs to blend with your background colour.

With the icon prepared, I did the bar chart for the breakdown of the quantity of pesticides banned in each of the three regions. I chose yellow because it denotes danger, caution - much used when referring to bio-hazardous substances.

With both the chart and the icons placed in your dashboard, you only need to place the images floating on top of the bar chart and adjust the size of the bars to fill your shape's transparent area. It is that easy, but it takes a bit of fiddling:

Tip: use the coordinates panel to make these adjustments, as it's much easier to place everything in the right pixel and keep the icons aligned.

5. Testing it with a non-data viz person:

This time I showed it to a friend. His first reaction was to say it looked cool and that it caught his attention - which is what I was expecting! I showed him the original viz as a reference, and he said: "oh if I saw that I'd just scroll down and wouldn't even bother reading the numbers..."

After looking closely at it, he asked if I could get details about which pesticides were banned in the US but used elsewhere as well! I guess we can all agree this is a big thing people generally thought about the paper, the original viz and the data provided for the week.

Weirdly enough, no comment was made about the design itself.

I asked, and his comment was: "oh, I can see we ingest a lot of poison unknowingly, and no one seems to talk about this, because it's too technical, but I think the poison bottles made me think about it".

Jackpot!

Feedback, feedback & more feedback

I always love to see someone who's not into data visualisation looking at one. The perception of what "works" is often very different than the one from a seasoned professional who has seen and worked on hundreds of visualisations before.

Not surprisingly, I got mixed reviews on this one. While on Twitter and with friends I got very positive responses to the visuals and they got the message straight away, it was brought up as a non-best practice example by the Makeover Monday team during this week's Viz Review.

I understand the commentary, and I agree it's not fit for all situations. But as all thing related to design, it comes down to purpose: mine was to make something eye-catching and try something new. It works.

Iterate!

This week's feedback was more of a general feeling than specific points to improve (the way I understood it).

The feedback:

• It was hard to see the numbers at the bottom of the poison flasks because they were just as small as the text beside it.

• The yellow colouring didn't contrast enough with the background to call attention to the most important numbers.

This is the final version, with the feedback implemented. You can also check it out here on Tableau Public.

What I've learned:

• My take is that data has a lot of parallels with language. You get a lot of signs you have to put together in a way that makes sense and tells a story - except that the story is a visual representation of data. In some instances, it will be like writing poetry - complex and full of intricate meaning. The reason why a writer cares about their audience is the same one why we should care about who our visualisations are aimed to: if they don't get it, you won't get engagement.

• Polemic feedback is always rich feedback. There's a lot to learn when two audiences have entirely different perceptions of something you put forward. Don't ever feel intimidated if you try something new and it doesn't quite land as expected!

• Data visualisation, like any science, is 1% inspiration and 99% perspiration. Iterate, try again, ask again. The more times you try something new and the more feedback you get, the better you'll get.

• The #MakeoverMonday community on Twitter is a great controlled environment for learning and experimenting with data visualisation features you wouldn't otherwise be able to. Often clients or stakeholders won't allow space for your experimental ramblings - at least you have a friendly group of people invested in it with you to share your ramblings with.

---

You can follow me on Twitter @DataRocksNZ and check my upcoming Makeover Monday submissions.

And while you're at it, check my Tableau Public Profile.