Now, almost a decade into my data visualization design journey, I have begun to question this practice and consider what happens when designers make the complex appear simple. In this article, I discuss the problem of oversimplification in data visualization and show how it can obscure (data) complexities that should be communicated.

The problem of oversimplification

We live in a complex world, and the visualizations we design should accurately and empathetically represent and celebrate the phenomena being explored (or explained). However, many practitioners in the field argue that the primary goal of data visualization is to present data in a way that is easy to understand and interpret. Now, such a goal seems not only productive but also laudable. But there is a seedy side to simplification that can result in designs that conceal important information, encourage overgeneralizations, and constrain creative expression. 

Giorgia Lupi, an information designer and data humanist, explores the theme of oversimplification, among others, in her article Data Humanism: The Revolutionary Future of Data Visualization for PRINT. She argues that, for many, part of the allure of visual design and, by extension, data visualization is its ability to simplify data. Indeed, there is something to be said for being able to reduce data into easily digestible visual representations. But do not be fooled by a designer’s “effortless” ability to make data look simpler than they are. Every decision a designer makes is deliberate and influences how their audience perceives the data and the real-life stories the data represent. 

Looking at several examples will help to illustrate this point. For brevity, I focus on three types of oversimplification: data aggregation, chart choice, and artistic license.

Data aggregation

Data aggregation is the process of expressing data in a summary form. When aggregating data, choices are made about what data elements should be minimized, emphasized, or removed altogether. Although data aggregation often occurs before a designer gains access to a dataset, data visualization designers and developers are increasingly being tasked with cleaning and preparing datasets before analysis begins. 

Thoughtful aggregation can make data easier to analyze—for instance, when your data are too granular or large to answer a question. However, if not approached with care, data aggregation can severely limit your ability to meaningfully make sense of your data. As an example, consider the case of “Underrepresented Minorities” (URMs) in Science, Technology, Engineering, and Mathematics (STEM).

According to the National Science Foundation’s National Center for Science and Engineering Statistics (2021), URMs are persons from groups whose representation in science and engineering education or employment (in the United States) is smaller than their representation in the United States (US) population. This includes individuals who identify as

• Black, 
• Hispanic,
• Latinx, or
• American Indian or Alaska Native.

Say you are a newly appointed Vice Provost for Institutional Analysis and Planning at a university in the US. Your first goal as Vice Provost is to better understand the state of racial (in)equity in undergraduate engineering degree completion. You ask your lead analyst to calculate degree completion rates for 2020-21 by race and ethnicity for the following groups: White Students, Asian Students, and URMs (which is standard practice). Your analyst provides the following summary information:

• URM Completion Rate: 54 percent
  (920 out of 1,700 Black, Latinx, and Indigenous students graduated on time in 2021)
• Asian Completion Rate: 76 percent
  (1,520 out of 2,000 Asian students completed their degrees on time in 2021)
• White Completion Rate: 71 percent
  (2,130 out of 3,000 White students completed their degrees on time in 2021)

The analyst also points out that when they performed a quality check, they noticed disparities in completion rates by individual URM groups. Curious to learn more; you request that information as well. After generating a report that presents degree completion rates disaggregated by all available race and ethnicity groups, a slightly different picture emerges:

• Black Completion Rate:  40 percent
  (200 out of 500 Black students completed their degrees on time in 2021)
• Latinx Completion Rate: 60 percent
  (720 out of 1,200 Latinx students completed their degrees on time in 2021)
• Indigenous Completion Rate: N/A
  (The university did not graduate any Indigenous students in 2021, nor do they currently have any Indigenous students enrolled.)
• Asian Completion Rate: 76 percent
  (1,520 out of 2,000 Asian students completed their degrees on time in 2021)
• White Completion Rate: 71 percent
  (2,130 out of 3,000 White students completed their degrees on time in 2021)

Here, creating an aggregate “URM” designation that groups Black, Latinx, and Indigenous students together masks variability in completion rates between students from different backgrounds. If the Vice Provost had decided to use data from the first report produced by the analyst, they would have never known that the engineering department did not graduate any Indigenous students during the 2020-21 school year AND that the Black student completion rate is substantially lower than the completion rate for Latinx students. 

By combining data for different subgroups into one larger one, a false sense of understanding is created about all students who are lumped into the “URM” category. This, in effect, erases the diversity of experiences, perspectives, and (potential) needs of those students. Indeed, by that line of reasoning, one could make an argument for further disaggregating the “Asian” category so as not to marginalize those Asian ethnicities that are often overlooked or not prioritized in conversations about STEM equity. 

I want to be clear that careful data (dis)aggregation is not a means to an end but a step in the sensemaking process. And the process of (dis)aggregation should not occur in a vacuum, nor should it be understood as static. Rather, it should be a (context-specific) liberatory practice that facilitates more comprehensive data stories and creates opportunities for differentiation and insight.

Now, I will turn to discuss chart choice.

Chart choice

Choosing the best visualization type is a challenge all designers face. Fortunately, there is a long history of research (e.g., Cleveland & Robert 1984  (paywalled), 1985  (paywalled); Pandey et al. 2015  (paywalled)) showing that certain types of charts and graphs are easier for audiences to understand. However, a designer can also choose a chart—intentionally or irresponsibly—that conceals data complexity or implies a misleading pattern or trend. Eli Holder and Cindy Xiong’s (2022) recent study explores how design choices can inappropriately convey and reinforce discriminatory messages. 

Through four experiments, Holder & Xiong (2022) show how visualization design can influence viewers’ perceptions about the subject presented. More specifically, they found that study participants were more likely to attribute differences in (social) outcomes to personal characteristics (e.g., people with better outcomes work harder than people with worse outcomes) when presented with a visualization that hides within-group variability (like a bar chart). On the other hand, participants were less likely to agree that differences in the outcomes presented were due to personal characteristics when they saw a visualization that emphasized within-group variability (like a jitter plot). 

Let me offer an example to illustrate this critical finding.

For more than half a century, ethnic and racial differences in educational outcomes have been the subject of much debate (Coleman 1968 (paywalled); Jencks & Meredith 2011  (paywalled)). These educational disparities (like the test score gap) have largely been framed in Black-White terms, with many (falsely) interpreting these differences to mean that Black students are academically and intellectually inferior to students from other backgrounds. One question that is often of interest to parents, educators, and policymakers is whether there are differences in reading test scores between students who identify as Black versus those who identify as White. 

Imagine you are the lead educational analyst for a school district. Your boss, the Director of Research and Policy Analysis, wants to know whether there are differences in the third grade reading test scores of Black and White students in the district. You run the numbers and find that the average reading score for Black third graders in the district is 148.48, and for White students, it is 172.90. You produce the following (horizontal) bar graph and present it to your boss:

Your boss looks at the chart and offhandedly remarks, “Looks like Black third graders can’t read.” You pipe up and say, “That is not true.” Your boss turns to you and asks, “How do you know?” Being the amazing analyst you are, you also produced a jitter plot showing the distribution of reading scores for both Black and White students, where the full range of values (e.g., minimum, maximum, and mean) can be seen: 

Now, your boss sees the fuller picture and realizes that, yes, on average, Black third graders scored lower in reading than their White counterparts. However, some Black students scored higher and others lower. 

Bar charts conceal how spread out a dataset is and overstate the appearance of differences between groups. And while their simplicity and widespread familiarity make them easy to digest at a glance, bar charts can be an irresponsible choice for presenting quantitative data that are grouped into discrete categories, especially if there is considerable variability in the outcome being displayed. In other words, visualizations that reduce a dataset to a single number can have the unfortunate consequence of misleading audiences at best and reinforcing stereotypes and societal biases at worst. 

Although the implications of Holder & Xiong’s (2022) study are significant for research and practice, they should have a profound impact on how educators and experts approach training future data visualization designers. Choosing visuals that oversimplify a dataset is often the result of a lack of experience with data and knowledge of the visual expression of data. All designers should be exposed to and have a thorough understanding of the variety of ways data can be (re)presented and encoded. This includes design considerations beyond conventional visualization approaches and popular charts that are standard in analytics and business intelligence tools but do not always fully capture patterns or trends or are not well suited for telling complex stories. 

My goal here is not to convince all of you reading to replace your bar charts with a jitter plot. Rather, like Holder & Xiong, I hope to impress upon you the “duty of care” designers owe not only to themselves but also to the public when it comes to (re)presenting data.

One final topic I will talk about in this article is artistic license. And by artistic license, I mean creative expression.

Artistic license

Creative expression in data visualization is a touchy subject. Many believe that creative expression and data visualization are inherently incompatible and design choices that deviate from “best practices” should be banished to the realm of data (or computational) artistry. For instance, Stephen Few, an information designer, has written about this topic on his blog, Perceptual Edge, in a now infamous post titled Does Art Play a Role in Data Visualization?. In the piece, Few argues that we, as data designers, should use the term “art” when referencing data visualization with caution. And I agree. But where he and I differ is in our understanding of visual design conventions or what he calls “aesthetics.” 

According to Few, there is no place for artistic license in “effective” data visualization; the creation of visuals and graphics should remain an objective, science-informed endeavor. That said, Few does recognize the importance of aesthetics in data visualization, but only because the design conventions (i.e., what works and does not) he employs in his work are rooted in “scientific research.” But what Few and others who adhere to this philosophy fail to acknowledge is that nothing operates in a vacuum. Everything—including scientific research—is influenced by time, culture, and current understandings of how the world should work. So, who is to say that data artists and data designers who experiment with more artistic forms of visual communication are not simply at the forefront of scientific developments that will lead to new norms in visualizing (or envisioning) information?

The tension between creative expression (form) and usability (function) extends beyond the theoretical to the practical realm. Current data visualization practices encourage sameness—copies of what others have designed. This has resulted in some designers becoming de facto technicians, blindly following “best practices” without questioning or tailoring them to fit their specific needs. Now, I am not suggesting that we throw caution to the wind and create works of art that do not communicate a story or allow the audience to discover their own story. What I am arguing for is a data visualization design practice that relies on evidence-based techniques grounded in scientific research— that acknowledges different ways of knowing and being— but embraces creative expression, such as experimenting with different encodings and visualization types. 

An example will help to illuminate this point.

Children living in immigrant families continue to be a growing segment of the US child population. In 2018, one in four—or 18.4 million—children in the US were born in another country or lived in a family with at least one foreign-born parent. 

Say you work as an information designer at an organization focused on immigration policy in the US. Your executive director requests that you create a visual that allows viewers to see where most children in immigrant families live. Using Kids Count Data Center data from 2005 to 2018, you produce two tile grid maps (one map using 2005 data, the other 2018) and stitch them together in a GIF animation:

Original Design: GIF of Two Tile Grid Maps

A tile grid map is a popular and effective way to present geographic data while giving each region (or state, in this case) equal visual weight. But there are other ways to present trend data that do not require animated transitions between two static visualizations.

In keeping with the “tile grid theme,” your next design brings an interactive element (i.e., scroll bar) to the visualization as well as a small multiples area graph. Users can click on the arrows of the scroll bar or use the slider to “activate” a red vertical line that highlights the current year’s percentage. Each tile also has a label showing the state’s name and the current year’s percentage.

Alternative #1: Interactive Tile Grid Map with Area Graphs

While interactivity can increase understanding, designers should not assume that viewers will (or should) click or hover to make sense of the presented data. Further, because many data points are displayed, forcing interaction via a scroll bar, in this case, can become burdensome for people with mobility impairments. Not to mention, the interactive element adds a level of granularity without much value. In other words, even without a data value label (that changes with each year), viewers can still see how values in the outcome change over time and judge how much using the map legend.

For your final piece, you produce a static visualization that uses the tile grid map as a foundation and blends popular visual encodings with a non-traditional chart type

Alternative #2: Tile Grid Map with Radial Column Charts

Here, a radial column chart is placed in each tile, and individual bars on said chart(s) represent a single year in the dataset, ranging from 2005 (one o’clock position) to 2018 (twelve). Instead of displaying values for specific years, each chart has a series of concentric rings representing 10 percent, ranging from 0 percent (innermost ring) to 50 percent (outermost ring). Moreover, the length of each bar is proportional to the percentage it represents, and viewers do not have to rely solely on bar length to decode the visualization. Color is strategically used to highlight and define regions with more (or fewer) children living in immigrant families.

My goal in discussing creative expression is not to suggest that all designers should aim to be more “artistic” in their approach or that visualizations with an “artistic” aesthetic are inherently better. Rather, the simplified setting of the example underscores how a dataset can be visually (re)presented in different ways. Some designs may resonate with your audience. Others, not so much. Experimentation, however, offers an opportunity to thoughtfully examine our design choices and decision-making and create more meaningful visualizations that will help your audience uncover new knowledge or come to a new understanding of the presented data.

We are at a critical juncture in the history of information design. Rising interest in data and growing familiarity with the tools and skills needed to present data offers the opportunity for, as Lupi puts it, a second wave of visualization that is experimental, unique, and “connect[s] numbers and graphics to what they really stand for: knowledge, behaviors, people.” This article only scratches the surface of the issue of oversimplification and the influence it can have on how we—as designers—choose to visually (re)present life. I will leave you with a quote from an essay titled Seeing Your Life in Data (paywalled), penned by Nathan Yau, a statistician and data visualization expert. One line in the piece perfectly captures my attitude towards the current moment and perhaps was (at the time) a foreshadowing of what was to come, “Data often can be sterile, but only if we present it that way.”

The post When Oversimplification Obscures appeared first on Nightingale.

Hello! I’m a student, data journalist, and designer with a love for all things dataviz. In accordance with annual traditions, I spent a lot of time thinking about my New Year’s resolution this year. Because the COVID-19 pandemic has kept me confined to my family’s home since March, I wanted to take more control of my everyday life and learn more about myself in the process.

I was inspired by Dear Data: a personal data collection project that Giorgia Lupi and Stefanie Posavec conducted for an entire year. In March, I did a three-week version to track the immediate changes to my communication style after evacuating my college’s campus. I decided to follow in Lupi’s and Posavec’s footsteps and collect a different piece of personal data every week for a year, but since I don’t have a partner, I added a twist.

Each week for fifty-two weeks, I’ll collect a different data point about my habits, activities, or attitudes, but those data points won’t be chosen at random. Fifty-two can be evenly split into thirteen groups of four, so each chunk of four weeks will have a theme. These themes can be anything: from entertainment to energy or color to connection. The important part is that those four weeks of data collection will be connected and (hopefully) reveal a greater truth about my relationship to that topic, which I’ll be able to share!

For the last four weeks, I’ve been collecting data about food. I started by tracking when I eat. I simply wrote down the time that I started and finished eating every meal. I also added little notes next to the time if there was anything that directly impacted the timing of my meal.

Because I recorded specific times instead of just how long I spent eating, I was able to organize this data visualization just like a typical hourly schedule. This layout revealed a few things immediately. First, dinner is by far my most consistent meal of the day. I’ve never been one to skip meals, but I know that I often prioritize doing work or attending meetings over having a meal at a “normal” time. However, since I eat dinner with my entire family, it’s at approximately the same time every night. My lunches show the opposite trend — they vary widely both in timing and length, likely because I am most busy during the middle of the day. I tend to just run down to the kitchen for a quick bite between meetings or working on assignments instead of taking the time to have a heavy meal.

Another thing I noticed is that I don’t snack as often as I’d imagined. I thought that I ate at least one snack a day, but during the week I recorded my mealtimes, I only snacked on three days, including one day when much of my eating was fueled by stress.

Moving into week two of data collection, I wanted to see in more detail what a typical week of meals looked like. I anticipated a relatively wide variety of dishes, as I do my best to eat a healthy mix of foods. I also challenged myself to be a bit more illustrative with my visualization, so I decided to draw simplified versions of each of my meals. These little drawings were made possible because I wrote down what I ate and the main colors of the foods, as well as how much I enjoyed the meal and how satisfied I felt when I was done on a 1–5 scale.

Overall, my meals were more monochromatic than I would have liked, though this may be in part because I standardized the size of the sides for aesthetic purposes instead of making them proportional to the meal that I had. Despite this flaw, the format of this visualization did allow me to size the “plates” of food by how large the meal was. Large plates primarily went along with dinners, while medium and small plates encompassed breakfasts, lunches, and snacks.

The data for week three were by far the most challenging to track. I tried to write down every conversation I had with another person that involved food, including the topic and participants. This was difficult because I live in a house with my two parents and five siblings, which adds up to a ton of conversations! I found myself tracking conversations in my head as they flowed organically and writing them down when there was a break or lull instead of trying to take notes in real time.

As is apparent from all the solid circles on this visualization, the vast majority of my conversations about food were with my family. Occasionally food would come up while chatting with a friend or neighbor, but because I cohabitate with my parents and siblings, I’m far more likely to talk to them about meal plans, how something tastes, or simply what’s in the fridge.

I also noticed that the blue rings, which symbolize conversations where someone gave their opinion about a food, tended to cluster together. Two big clumps of blue rings can be seen on January 16 and January 19, which coincided with a baking itch that I decided to scratch. On each of those two days, I baked a different kind of pie and most of the blue rings were compliments about my handiwork!

After reflecting on my first three weeks of data collection, I chose to make my last week of food data collection slightly more reflective. During week four, I wrote down all of the thoughts I had about food and when I ate within the web of those thoughts. This manifested into a noticeably different type of visualization than the first three weeks.

When I examined my week four data to make this visualization, I realized that all of my food-related thoughts could be placed into four categories: feeling hungry, craving a particular food, thinking about my next meal, or telling myself that I should take a break to eat. I represented each of these thoughts with a leaf on a stem. That stem can either be cut off by eating (a flower) or by ending my day (an unadorned stem). It’s also possible for me to eat a meal without actively thinking about it in advance, which is depicted as a flower in the dirt.

This visualization suggests that after I think about food two or three times, I usually decide that a meal is in order; however, there were occasions throughout the week where I thought about food five or even six times before giving in to my mind’s insistent messages. Once again, I noticed that all of my snacks came at the end of my days, as opposed to in between lunch and dinner, which is when my younger siblings typically insist on a treat.

These last four weeks of personal data collection have allowed me to examine my relationship with food in a unique way that revealed some surprising trends. I’m not as much of a snacker as I thought I was and I eat fewer greens than I had imagined. I’m also pretty responsive to my body’s signals for food, which was a welcome revelation!

With the first four weeks of data collection finished, I’ve already begun the next theme: movement. I’m updating my personal data collection journey weekly on my Twitter, Instagram, and TikTok, so if you’re interested in following along with my visualizations in real time, look no further! In the space between this article and the next, I invite you to follow along or even try some data collection of your own and share it with someone. Until then, stay tuned!

Emilia Ruzicka studies data journalism at Brown University and will graduate in May 2021. Her current work includes a year-long personal data collection project, a podcast about the USPS, and her senior thesis. Find out more at emiliaruzicka.com.

The post Food for Thought: Part 1 of a Yearlong Personal Data Project appeared first on Nightingale.

This gave me an idea for entering a local data storytelling challenge a while back. I would explore the profile of popular wedding songs in a data visualization.

Brainstorming

There are several ways to visualize music, but I was rather inspired by Susie Wu’s film flowers and Amy Cesal and Zander Furnas’s wedding guest badges, which were in turn inspired by Georgia Lupi’s data portraits for 2017 TED Talk attendees. In their respective works, data points on a given subject were turned into a playful cocktail of shapes, colors, and symbols.

I knew that the Spotify API provided information on a song’s audio features such as its duration, energy, and level of cheerfulness. Putting the two together, I decided to visualize individual songs as a combination of their respective features. Tackling how was left to a later stage.

Clarifying objectives

Confession: I find it’s easy to get distracted or lost when working on an open-ended project like this. To help anchor the project, I came up with some key guiding questions. Contrary to how it may appear in this article, the process was not a linear one. Rather, I refined the questions as I poked around the data and designed some prototypes. These questions ended up being:

• Why do certain songs keep appearing on wedding playlists? This was the big, burning question that motivated the project in the first place.
• What features do wedding songs tend to have in common? How do these differ from other popular songs? These were the supporting questions to guide the data exploration.
• What does our choice of music features ultimately say about us as humans? More specifically, how do we recreate memories or set the mood we want for weddings? These were the probing questions that reflected the takeaways I’m hoping to get at the end of the project.

Data exploration

Initially, I had the thought to compile my own list of wedding hits by scraping every Spotify playlist with “wedding” in the title and seeing which songs appeared most often. However, this approach proved to be rather cumbersome, so I turned to a shortcut: Spotify’s declaration of the 50 most popular wedding songs in the form of their own playlist, “Most Popular Wedding Songs Globally.”

We often need to have a comparison or benchmark as a context to get a meaningful sense of the data. For example, what does it mean if a wedding hit has an energy score 0.6 out of 1.0? Is this typical of songs in general or just wedding songs? For comparison, I gathered songs from two other Spotify-curated playlists, the “Top Tracks of 2018” playlist for popular songs and the “Broken Heart” playlist for songs that I felt would be the polar opposite of wedding songs.

For each feature (e.g. energy, mood, etc.), I charted the distribution of each group of songs. One thing that stood out was that while wedding hits in general tend to be more cheerful than heartbreak songs, two types of wedding hits were rather common: super upbeat songs and rather heart-rending ones. Also, it seems that while wedding hits tend to be more acoustic in nature than top tracks (which were mostly electronic dance music or hip-hop), they were less so than the heartbreak tracks. For those curious, Spotify’s glossary on the audio features is available here.

In terms of genre, unsurprisingly, the popular wedding songs were more often pop or soul music. Spotify’s records of song genres were mostly missing so I got the data by web scraping Google’s featured snippets for genre search results. Not exactly a foolproof method, I know, but it sufficed for providing an overall sense.

Design process

As I was working with wedding hits, I thought it would be fun to represent the songs using favors commonly found at weddings.

The process of imagining and defining the visual concept started with rough sketches on paper. I found that cupcakes provided the sweet spot as a canvas that could accommodate a layering of design features without being too complex. Mapping the music features to cupcake decorations required further rounds of experimentation to see what worked and what didn’t.

While checking online on how I could automate design patterns based on numerical properties, I came across mosaic art and spirographs.

Creating the desired spirographs took many iterations. Special shout-out to Joel Schneider’s awesome spiro package in R! Other aspects of the design were done directly in Inkscape, an open-source alternative to Adobe Illustrator.

Baked cupcakes

I pulled everything together in an infographic for the data storytelling challenge. To my delight, my submission was awarded a commendation for its creativity and design.

As shown in the infographic, it seems certain types of music do make popular wedding songs. Soulful, slow-tempo ballads (perfect for those reflective moments) occupied top positions in the ‘Most Popular Wedding Songs Globally’ playlist. High-energy music, which are great for dancing to, made it to the playlist as well. Looking at the song titles, it also appears that people tend to lean toward music that is overtly about romantic ideals. See “Marry You”by Bruno Marsor “Marry me” by Train for instance. However, the fact that “Hey Ya!”by Outkast—a deceptively depressing song about relationships— is also a wedding hit suggests that people pay more attention to the vibes in the music than the actual lyrics.

Looking back, there are definitely a few things I would have done differently. One is to make the interpretation of cupcake images more user-friendly, perhaps with smarter labeling or an easier legend.

All in all, it has been a great hands-on learning journey!

The post Sweet Love: Popular Wedding Songs Reimagined As Cupcakes appeared first on Nightingale.

I know that I am not the only one that, when listening to music, likes to imagine shapes and colors as well as growing and ever-evolving visual patterns. Up until recently, I did not know that music visualization¹ ² was a common thing in creative coding projects. As an attempt to reconcile my neuroscientific and academic background with my current data processing environment, I created an audio and music visualizer that uses the information within audio streams to create a neural forest, all in real time. Here is how it looks after a fragment of “Claire de Lune” from Debussy:

NOTE: I am not going do describe detailed mathematics here. If you are interested in knowing more about the algorithm and how it works feel free to contact me! Source code can be found at the end of the text.

The sound Neuron forest in detail

Looking at an original drawing by 1906 Nobel Prize winner Santiago Ramón y Cajal, we can see that a pyramidal neuron looks like this:

The soma in the center (the body of the neuron) has an axon (the wider branch) and a dendritic tree (rest of the smaller branches)³. Through a process called neurogenesis⁴, these specialized cells grow from simple immature somas to intricate mature neurons⁵ like the one we see here, with branches randomly spreading out of the middle soma. If we could simulate this growing then, it seems that a good idea to do so is by means of random walkers⁶, wherein very general terms and without going into mathematical details, the position p of every growing branch b at the step t+1 is randomly chosen. True random walks will generate very intricate dendritic webs (there is no directionality gradient) so it would be wise to change direction based on a probabilistic approach, where all new positions in branching patterns are only changed if that given threshold is passed. Very importantly, we can see from Cajal’s original drawing that the farther away a dendrite or axon goes from the central soma, the thinner it gets and the more intricate branching patterns emerge. We can also replicate this again with a probabilistic take, where after a certain axon/dendrite width is passed, we generate a new branch (or walker in mathematical terms). If you are interested into the mathematics of random walks, which are applied in the code, there are many sources that go into more details⁶ ⁷.

To achieve all this, I used the Java-based programming language Processing. Processing has some very powerful tools and functions that allow it to integrate graphical and audio streams into one single pipeline, making the job easier.

Luckily for me, the Processing community is very active and through openprocessing.org I found a sketch code simulating a growing tree that suited my needs. The original random branching algorithm, to which I adapted and added sound input (and more) to create this project, is taken from here.

After some coding (a link to it at the end of the text), I managed to create simulations that imitate neuronal branching. Compare a drawing of a neuron using the code with what Cajal drew:

Similar, don’t you think? Interestingly, given that the regularity is provided only at the initial branching parameters (like size, reach and transparency) while branching follows a probabilistic random growth, each time you run the drawing you will get a completely different neuron:

With this, we have the base of the drawing figured, now let’s bring this neurons alive with sound and color!

Remember when I said that the drawing parameters are fixed? What about changing these parameters based on an input? Even further, what about making this input audio, so depending on the input, the growing pattern and velocity changes? This is exactly what I did next.

The easiest approach with sound would be to work with a real-time amplitude (or intensity) stream of the signal as a growing parameter. In very simple pseudo-code this would look like this:

sound_neurons(neuron_width,neuron_reach,amplitud_stream){ drawing_commands;}

Where amplitud_stream is the only parameter that is updated via real-time audio streaming (see source code for more information). First, let’s change the diameter and transparency of the soma and dendritic branches based on this stream at the moment of running the simulation. The louder the input, the larger the neuron will be. Also, let’s track the drawing so we can see how the neuron branches in real time.

Now things start to look very interesting! With this tweak, we are beginning to have sound-reactive neurons. With a quiet input (which was me literally whispering into my microphone) I got a small neuron. After a very loud input, the next drawing is much bigger! This is why I decided to call these drawings sound_Neurons.

Another thing we can try is to manipulate the drawing speed, again with the streamed amplitude. Look again at the cover image of this story and you might see that the growing speed in each sound_Neuron follows a bum bum….. bum bum…. bum bum… rhythm.

I then decided to randomly change the colors (in HSV space⁸ in the video example) of the neurons each time they appear so the final picture is more colorful:

Finally, Processing is able to stream and read audio files, so the final and logical thing for me to do was integrate these colored sound_Neurons with music! For this, I decided to use “Claire de Lune,” the beautiful piano suite from Claude Debussy, as the driver for a neural forest. By driver I mean that the intensity and changes in music rhythm will dictate how this forest grows. The final result is this image:

Generated in real-time in the video at the beginning of the text while using a dark background so neurons really fill the space with sound. Even after many runs, I still found it mesmerizing to watch. I can run the simulation many times, over and over again.This side project of mine turned out to be much more than what I anticipated in the beginning. Listening to music through these neural forests helped me bring to life the shapes and colors I talked about imagining in a rich and rather personal way. I like to think of this as my take on trying to mix a little bit of science with art and signal processing. Like so, this project is a reflection of who I am. I did my graduate studies in Neural Networks and as someone who loves music and visual arts, this was a mean of expressing my passion for both sides.

I hope you enjoyed reading and watching these animations too. I certainly hope to continue exploring this beautiful field of music visualization in the future and with luck, again with sound_Neurons.

NOTE: What song would you like to see colored by neurons? Let us know in the comments and we will randomly select one or two and generate a new sound_Neuron forest.

Thank you for reading!

Source code here

References:

[1] https://en.wikipedia.org/wiki/Music_visualization

[2] https://medium.com/nightingale/data-visualization-in-music-11fcd702c893

[3] https://qbi.uq.edu.au/brain/brain-anatomy/what-neuron

[4] https://qbi.uq.edu.au/brain-basics/brain-physiology/what-neurogenesis

[5] Kempermann, G., & Overall, R. W. The small world of adult hippocampal neurogenesis. 2018. Frontiers in neuroscience, 12, 641.

[6]https://www.mit.edu/~kardar/teaching/projects/chemotaxis(AndreaSchmidt)/random.htm

[7] ttps://medium.com/@ensembledme/random-walks-with-python-8420981bc4bc

[8] https://en.wikipedia.org/wiki/HSL_and_HSV

The post Drawing Neurons From Sound And Music In Real-Time appeared first on Nightingale.

Walking and birdwatching is normally my way of relaxing, getting away from the laptop and clearing my mind. It is also my way of getting creative. Coming up with ideas and solutions for projects that I may be working on or for potential new graphics.

Looking through some of the books produced in the 60s and 70s, I came across some fantastic black and white hand-drawn graphics in the ‘New Naturalist’ series I have ranging from birds to moths to hedges. Each book contains a single subject and hence the scope of the graphics is limited to that. However, you can see that the expert in each case is really trying to share their knowledge with the reader in a visual and understandable way.The graphics range from simple small multiple distribution maps of the UK’s birds of prey (as seen above) to line charts showing egg thickness changes due to pesticide use. More complex graphics range from network charts through to illustrative drawings (anatomy of a dragonfly nymph — the stage of a dragonfly and damselfly that lives underwater).

Flicking through ‘Dragonfly’, a complex looking graphic stands out showing the activity pattern of a male dragonfly. It’s striking because of its shape. It looks like a pyramid drawn within concentric circles. An assortment of wavy lines and dots complete the graphic. Because of the complexity, the author has captioned the graphic with detailed ‘how to read the graphic’ text (something we should all think about when producing new, innovative or complex graphics).

When you take a closer look it’s actually a complex data visualisation based on a 24-hour clock. Showing time, position (is the dragonfly near water or away from water), activity (is it perched, cleaning or hunting) and reproductive behaviour (mating or fighting) all portrayed within the graphic. Very smart, and once you have read the caption, very clear! The more I look at this the more I admire it! All this is hand-drawn and only using black and white line work with various types of iconographic shapes.

Another from the series of books looks like a square pie or treemap type of graphic with pie charts overlaid! Sounds bad but have a look. Again it comes with a good graphic explanation. It is based on a 4-acre square field surrounded by hedges. The pie charts show the placements of traps used to catch field mice and bank voles at various times in the year. You can clearly see that, whatever time of the year, the mice can be caught everywhere, whereas the bank voles kept primarily to the hedges (edges). A clear, simple, useful black and white graphic.

In both of these examples, if done today, I would suggest a good headline and explanatory sub-text would be a useful addition, rather than the accepted caption at the bottom (something that scientific papers still need to think about). Narrative text plus pointers to areas of the graphic, but really, I think these are fantastic visuals that get the point across to the reader really well. Simple and effective. Paring back the graphic to what is needed without distraction…obviously helped by the use of only black and white and shading.

There are many more examples across the books that show the creativity needed using just back and white. You can see a couple below, all simply explaining with minimal distractions.

Inspiration and creativity can come from anywhere in life. It’s not just about looking at what is being produced now looking back at what was being produced last year, or 10 years ago (see my blog looking back at what I was producing 10 years ago http://nigelhawtin.com/10-years-ago) or even further back is so important. There have always been constraints, whether that was having to hand-draw things, lack of colour, knowledge or time.

For me, creativity and inspiration can come from the peace and solitude of walking in the countryside at 5am taking in all this world has to offer, looking through old books and magazines, traveling or finding something online that takes my interest.Look up, look down, look to the future, but don’t forget to look to the past as well — inspiration is all around us.

The post Finding Inspiring Data Visualisations In Mid-century Nature Books appeared first on Nightingale.

Enter the sticker.

Like trophies, stickers can celebrate victories, but because they’re inexpensive to produce they commemorate a broader category of achievements.

Stickers celebrate wins large and small. Awarding stickers to kids in the classroom is so common that “you get a gold star” has become an expression of approval for a job well done. / Girl: César Rincón (license)

A numbers game

Trophies can recall happy memories, but a large trophy collection has a quantitative message. A row of trophies doesn’t spotlight specific achievements. Instead, it encourages the viewer to think about the scale of the collection. It’s a visualization of winning.

Because stickers are plentiful and don’t require a lot of space to display, adhesive trophy collections are not hard to find. The buckeye leaf stickers that adorn Ohio State University helmets are well-known among fans of U.S. college football. The criteria for earning these stickers have evolved over time, but in general they’re awarded for strong performances in a game. Here, the specific achievement associated with each sticker disappears completely. All that remains is the count.

A suitcase covered with old hotel luggage labels is an iconic image of travel. It also happens to provide a visualization of how well-traveled its owner is, for the benefit of other passengers on a train.

A more modern equivalent: the stickers on the side of this recreational vehicle (RV) allow the owners to document the states they’ve visited and, it appears, to note a few activities along the way (square dancing in Nevada, whale watching in Washington).

Location adds meaning

Large trophy collections need space. Where a trophy collection is housed often asserts ownership of the reflected glory: a high school, for example, or this fire station.

On the other hand, sticker trophies don’t take up much space, and they gravitate toward places or objects that played a role in their stories, adding some qualitative color to the quantitative picture.

The locations of the stickers above do suggest ownership, but they also provide additional context about what the stickers mean or how they were acquired. For example, the RV map suggests not only that its owners have visited 18 states, but specifically that they have done so in that vehicle, evoking cultural connotations that come along with that mode of travel. The location of the map also reinforces the explicit depiction of the 48 contiguous U.S. states as the likely priority for destinations. (I did find some other sticker maps that included insets of Alaska and Hawaii, causing me to wonder how many people actually choose to transport their RVs to Hawaii.)

This guitar case also tells a specific story. Like the suitcase above, it features travel souvenirs (I see stickers related to tours, concerts, and radio stations, among others) but the fact that these are plastered on an instrument case suggests that these are places its owner visited as a performer. I don’t know this to be true, but it’s the story that emerges.

Group efforts

The sticker collections above reflect the achievements of individual people, but trophy collections often represent the efforts of many. Those on display in the fire station photo above were collected over time (the years 1983 and 2001 are legible) and are sure to represent wins by people who never met. The common thread is the station.

So-called “victory markings” commonly adorn military planes, a practice dating to World War II. These are usually painted stencils, not stickers, but they serve the same function. Sometimes the pilot’s name is displayed along with each marking, but the linkage to the event is with the plane.

Finally, I found a crowdsourced collection in these photos of astronauts bound for the International Space Station. Crew members are shown affixing mission insignia stickers to the wall of the plane that transports them to the launch site, building a visualization over time.

This accumulating collection of stickers is a little different than the others, and it may be my favorite. I’m never going to be an astronaut (sorry, eight-year-old me) but there’s something about this improvised ceremony that captures the essence of it. Working aboard the ISS is a milestone for everyone involved, but this humble composition of faux wood and paper on what looks like an ordinary passenger jet is not a grand display for the world to behold. Rather, it’s a chance for those few who stand before it to connect with those who came before and those who will follow, to say “we were here,” and to make their small mark on something much larger than themselves.

Thanks to Jason Forrest. 

The post Visualizing Small Victories appeared first on Nightingale.