top of page

Visual Evolution: A Moore's Law Makeover

Writer's picture: Zane WolfZane Wolf

In this post, I’ll walk through how I adapted author-contributed figures for an op–ed article into the Scientific American house style, from understanding the science behind the graphics to the various drafts that shaped their final form.


The Science Behind the Story

In 2015, the Laser Interferometer Gravitational-Wave Observatory, or LIGO for short, detected for the very first time a ripple in space-time. Einstein theorized that our universe took the saying 'the fabric of reality' a bit too literally. The combination of space and time weaved together tautly to form something akin to a trampoline. Drop a basketball onto the trampoline and the surface dips towards the center of the ball. Have a big large enough trampoline, and it probably only dips locally - the other side of the trampoline might remain nearly flat. If you have a tennis ball already on the trampoline when you add the basketball, chances are good that the tennis ball will roll towards the basketball. It, having far less mass, fell into the basketball's gravity well.


Now what if instead of a basketball and tennis ball, you have two black holes that are orbiting each other ever closer, both caught in each other's gravity wells. These heavy blobs of darkness are circling towards each other with such speed on the trampoline that they produce a bow wave in front of them (up) as well as a wake behind them (down) as they push across the fabric of reality. If you have trouble imagining how a bow wave is produced, place a tea towel or similar fabric on a table, and using your fingers push the fabric across the table. Did it scrunch close to your fingers and remain relatively flat and stationary on the other side? Same thing. (Now ask me how this explanation relates to how fish swim and airplanes fly.)


The cyclic nature of up down up down in the fabric as these two black holes orbit around each other produces waves that ripple outward across the fabric. As the black holes get closer to each other, they will orbit each other fast and faster and the waves will grow and grow eventually to the point where they will ripple far enough out into space that we can still detect their faint signatures even at great distances from the origination point. The detection of this faint ripple or 'gravitational wave', which is only apparent on a size scale of less than one ten-thousandth the charge diameter of a proton, strongly supports Einstein's theory of relativity and the space-time continuum.


The first LIGO was built in 2002 and ever since then, the technology invented to help us confirm the existence of these then-hypothetical waves has advanced. Similar to other technologies like cameras and satellites, each advancement allows us to be able to see farther and with more detail, increasing the chances of observing colliding black holes by detecting the waves they produce. By multiplying the volume of space that we can observe using LIGO by the number of possible merging black holes we expect to see within that volume, we can generate a number of possible black hole merger detections per year.


Dr. Imre Bartos and Dr. Szabolcs Marka wrote an opinion piece for Scientific American positing that the advancement of the LIGO technology essentially allowed for the number of possible detections per year to double every two years. This doubling, this exponential growth of a technology, has a name: Moore's Law.


Since Gordon Moore coined the law to describe the rapid advancement of transistor technology in the 60s, Moore's Law has been applied to a number of different technologies. One important aspect of Moore's Law, though, is that it predicts a doubling of the technology every two years without a corresponding increase in the finance needed to fund that advancement. Being able to say that LIGO follow's Moore's Law, that it can advance rapidly without needing an exponential growth in budget to sustain it, might make procuring future funding easier.


In the House Style


In support of their argument, Marka and and Bartos included three figures in their article draft, two of which the editor, Lee Billings, decided would be good to include in the final version. They are reproduced here with the authors' permissions.


The first figure illustrated the exponential growth of possible detections per year with each subsequent improvement of LIGO, and the third showed the five-fold increase in the number of authors on LIGO papers throughout the years as a supplementary measure of the how important the gravitational wave field has become and currently is. (The abandoned third figure showed funding over time—it greatly resembled the increase in authors over time.)


Each figure is color-coded for eras. The number of detections per year are color-coded by the improvements made to the LIGO detection facility, after which LIGO received a name change to demarcate the improvements: initial LIGO, Advanced LIGO, LIGO A#. Cosmic Explorer will be an entirely new facility.


The eras in the line chart show collaborations. Virgo and KAGRA are both gravitational-wave detection facilities in Europe and Japan, respectively, that formed collaborations with the LIGO facility to improve detection capabilities.


I came on to remake these figures in the 'house style'. This blog walks through the redesign process of each of these figures, how I went from the originals to the final versions you see in the published article.


It took me a long minute to understand what my editors meant by 'remake in the house style.' This was my first assignment where I did not design and conceive of the graphics in the article from start to finish, but had to work with someone else's figures and make them ready for publication at Scientific American.


At first, I interpreted my assignment literally—remake the graphic using our color palette, fonts, and sizing specs. Unsurprisingly, my first round drafts looked much the same as the original. Go figure.



But when I got first round comments back, I quickly realized that this is task, "remake in the house style," meant more than tweaking the components to meet our style guide (though that's important, too). These original figures were made for an academic audience, and I needed to reimagine them for a general audience.


It's like something clicked in my head at that point—this is what it means to be a graphics-focused science communicator. As a run-of-the-mill science communicator, I can interpret and translate the science and concepts in an article from jargon and technical details to formats that the general person might be familiar with or find more approachable - like trampolines. But, this was my first time doing that process for an article's graphics.


With that realization, I started thinking about what changes I would make to the originals. The point was never to 'improve' the original figures. As both a former academic and someone learning data visualization best practices, I feel confident saying there is nothing wrong with them to begin with. In fact, the first figure is a type of chart I don't think I've seen anywhere else before. I've been calling it the scatter-gantt because that is what is it—a combination of a scatter-plot and a gantt chart. I love it.


But what could I add in terms of clarity and story? What details weren't necessary to understand the main take-away? Were there other ways to display the data that would better support the authors' key points?


And with that, I got to work. For both of the figures below, I used Flourish to make the base charts and then Illustrator to add the different design elements. Funnily enough, remaking one of these figures gave me a day-long headache, and remaking the other was a tremendous amount of fun. Let's dig into it.


The Scatter-Gantt

These were the comments I got on that first draft:


I recommend stepping away from this and trying to shake off your previous knowledge. Then look at it with fresh eyes. What information does the reader need at what point in the graphic to understand what they're looking at here? I suspect the academic in you is trying to nudge the designer to the side a bit ;). Take a fresh naive look at this and question every label. What does it mean?

So I took a step back—too large of a step back, frankly—and went back to the drawing board. Literally. Was this chart the best way to visualize this data? If the point was that the possible number of detections scales exponentially, maybe I should show the difference between the projected numbers and the actual exponential growth line?





Residuals aren't exactly general audience savvy though. Observant readers might also realize that I plotted the exponential growth line differently. The authors plotted it starting with the 01 observation run in 2015, and I started mine in 2003, not understanding why the authors chose to start it in 2015 originally. After this, the authors and I discussed the data behind this chart so I could understand exactly what they were showing and why they chose to plot it that way. If you have questions about the data, just ask.


After ransacking my brain for other, traditionally-academic chart types that would fit, I had the realization that just because the original chart was done using a somewhat standard academic chart-type doesn't mean I was also confined to that category of visualization. This was the first time I used the 'Break the Box' technique by Alli Torban (I highly recommend her book Chart Spark).




But all of these other ideas of visualizing exponential growth had problems. Namely, abstracting the take-away and making it a little harder to see the exponential growth. Plus they would take up a LOT more space. Exponentially more, one might say.


So after this completely unnecessary brainstorming phase, I came back to the original design with the realization that there were probably much smaller-scale adjustments I could make that wouldn't require remaking the entire chart.




And so through a series of small changes - namely, coloring the points themselves rather than the background, directly labeling the different LIGO phases, and removing the exponential growth line, this chart became a little less visually overwhelming and more approachable for a general audience. Sometimes a light touch is better. And ultimately, I'm really happy with the outcome and that we kept the authors' scatter-gantt design.


Now, having said that, let's take a look at the second chart.


The Line Chart

In comparison to the scatter-gantt and the massive amounts of wrestling I undertook to end up essentially where I started, my process for bringing clarity and story to the line chart was straightforward and confidence-boosting.


It's been a while since the intro and to recap: this line chart shows the five-fold increase in authorships of LIGO papers since 2003 as a way to show the increasing importance of the field itself, and the colors show subsequent collaborations.






Once I realized I didn't have to stick to the original chart type in order to 'remake it in the house style', I had the immediate idea to use a waterfall chart instead to show both the overall increase as well as the incremental changes per year. It was my first idea and it felt so right that I didn't even bother brainstorming other options.


I started with basic annotations to help explain how to read the chart, and started with a poor green-red color scheme to show increases and decreases. I knew I wouldn't keep this color palette for two reasons, though. First, it's not color-blind friendly. Secondly, the red really stands out and makes the years where the number of authors decreased more important than they really are. It's the overall increase that is important.


Then I reorganized the format to put the years and collaboration labels at the top. Since these labels help orient the reader to what is happening, I wanted them to be read first. Unlike the scatter-gantt where it made more sense to directly label the eras and remove the colored backgrounds, here it made more sense to keep the labels justified to the axis to reinforce the successive nature of the eras. Categorical labels versus sequential, basically. The background colors helped tie the bars to the labels directly despite the distance between them.


And this is where I also started adding annotations to help explain the large changes seen in some years. I was at Georgia Tech in 2015 when LIGO first detected gravitational waves, and my two best friends actually worked with Dr. Laura Cadonati, the head of the GT LIGO lab. The first detection was a HUGE deal and all the School of Physics could talk about for days (probably weeks). I wanted to bring those paradigm-shifting moments to bear in this chart and show the impact they had on the number of authors.


By color-coding the bars by the eras, you can also directly see how the addition of the new facilities impacted the number of authors. It made more sense to put this observation/explanation in the caption than in an annotation, since they were a core part of the chart and not ancillary context like the detection events.


The annotations also have two different colors - saturated black for the annotations that help the reader understand the chart itself, and grey for the annotations that add bonus context.


And then lastly, I changed the color scheme from blue-yellow-red to yellow-blue-purple. The change from light to dark, less to more saturation, helped reinforce the linear change in numbers from less to more.


I struggled with label positions while remaking this chart, honestly. There aren't really hard and fast rules with annotation placements. Start with the cardinal directions - north, south, east, and west, and then rearrange as necessary. But how far away should they be, how close? What gets a straight leader line and what gets a curvy one? There's never really one right answer for these types of decisions—ask 10 other designers, you'll get 10 different placement options. In situations where there isn't a right answer, good enough is good enough.


Wrapping Up

Remaking these charts served as an incredible learning experience, if you couldn't tell by just how lengthy this blog is. I had a rocky start but ended with two charts that I'm incredibly fond of, and fond of for different reasons. What more could I ask for?


And bonus feedback from the authors:

It was nice to work with you and I had received very positive feedback on your graph designs, that I also liked very much. We are honored that you rely on these graphs in your portfolio, you should be proud of them.

0 views0 comments

Recent Posts

See All

Comments


bottom of page