Data Visualization by Jeffrey Dick

Tap an image to view the full-size version. Tap again to return to this page.

• Imagine that you are looking at a tree from above. The Starburst Plot represents the top surface of an evolutionary tree in order to see differences and overlaps between groups of species.

  publication · code · documentation


• Neutral pH is an important concept in chemistry, but no direct counterpart exists for oxidation-reduction (redox) reactions. The Affinity Slopes plot is used to find a midpoint on the redox scale; it is where reduced and oxidized organic compounds have nearly equal stabilities.

  website · publication · code


• The Stability Diagram reveals how minerals with different elements (bold text colors) dissolve in water (blue areas) under specific conditions (plot axes).

  website · publication · code


• Zinc 3D is looking through a stability diagram; the third dimension (z direction) is used to predict high-energy states. Check out the interactive plot.

  Show code

  # Zinc_3D.R
  # Make 3D solubility plots for Zn minerals
  # 20210309 First version jmd
  # 20240320 Added HTML export
  
  library(rgl)
  library(CHNOSZ)
  
  # Set plot resolution
  res <- 50
  # Load thermodynamic data without organic species
  OBIGT(no.organics = TRUE)
  # Set up system
  # - Zn first for swapping Zn-bearing minerals in solubility calculation
  # - quartz with default activity = 1 for quartz saturation
  # - H2S with default activity = 1
  #   (gives Stotal = 1 in mosaic calculation)
  # - CO2(gas) with fugacity = 10 for f(CO2) + f(CH4) = 10
  #   (CO2/CH4 ratio in mosaic calculation is controlled by fO2)
  # - Cl- and H2O with default activity = 1
  # - O2(gas) and pH as plotting variables
  basis(c("Zn+2", "quartz", "H2S", "carbon dioxide", "Cl-", "H2O", "oxygen", "H+"))
  basis("CO2", 1)
  basis("pH", 7)
  # Load Zn-bearing minerals
  species(c("sphalerite", "zincite", "willemite", "smithsonite"))
  # List aqueous Zn species
  iaq <- retrieve("Zn", c("S", "C", "Cl", "O", "H"), "aq")
  # Set up mosaic calculation to account for speciation of S and C
  bases <- list(c("H2S", "HS-", "HSO4-", "SO4-2"), c("carbon dioxide", "HCO3-", "CO3-2", "CH4"))
  # Calculate solubilities of minerals and overall solubility
  s <- solubility(iaq, bases = bases, T = c(25, 300, res), O2 = c(-60, -20, res), in.terms.of = "Zn")
  # Convert log activity to logppm
  s <- convert(s, "logppm")
  
  # Make 3D plot 20210322
  # For outputting to HTML (see ?rglwidet) 20240320
  save.HTML <- TRUE
  cex <- 1.5
  if(save.HTML) {
    cex <- 1
    save <- options(rgl.useNULL = TRUE)
    open3d()
  }
  
  # Get x and y values
  x <- s$vals[[1]]
  y <- s$vals[[2]]
  # Overall minimum solubility (most stable mineral)
  zmin <- s$loga.balance
  # Colors and mineral abbreviations
  col <- c("skyblue", "seagreen", "coral", "goldenrod")
  abbrv <- c("Sp", "Zct", "Wm", "Sm")
  dz <- 0.15
  zlim <- c(-1, 3)
  
  # Loop over minerals
  for(i in 1:4) {
  
    # Get z values for surface
    zlines <- zsurf <- s$loga.equil[[i]]
  
    # Get lowest values for surface (minimum solubility)
    zsurf[zsurf > zmin + 0.1] <- NA
  
    ## Plot surface (has issues on phone)
    #if(i == 1) persp3d(x, y, zsurf, col = col[i], polygon_offset = 1, xlab = "", ylab = "", zlab = "", axes = FALSE, zlim = zlim)
    #if(i > 1)  persp3d(x, y, zsurf, col = col[i], polygon_offset = 1, add = TRUE)
    # Plot lines on the surface
    #persp3d(x, y, zlines, front = "lines", back = "lines", lit = FALSE, col = paste0("light", col[i]), zlim = zlim, add = TRUE)
    if(i == 1) persp3d(x, y, zlines, front = "lines", back = "lines", lit = FALSE, col = paste0("light", col[i]), zlim = zlim,
                       xlab = "", ylab = "", zlab = "", axes = FALSE)
    if(i > 1) persp3d(x, y, zlines, front = "lines", back = "lines", lit = FALSE, col = paste0("light", col[i]), zlim = zlim, add = TRUE)
  
    if(i == 1) {
      # Plot sphalerite label just above highest solubility at highest logfO2
      loga <- zsurf[, 1]
      imin <- which.max(loga)
      ztxt <- loga[imin] + dz
      ytxt <- s$vals[[2]][1]
    } else {
      # Plot zincite, willemite, and smithsonite labels just below lowest solubility at highest logfO2
      loga <- zlines[, res]
      imin <- which.min(loga)
      ztxt <- loga[imin] - dz
      ytxt <- s$vals[[2]][res]
    }
    xtxt <- s$vals[[1]][imin]
    text3d(xtxt, ytxt, ztxt, abbrv[i], cex = cex)
  
  }
  
  ## Use this to get the rotation matrix interactively
  #userMatrix <- dput(round(par3d("userMatrix"), 3))
  # Set the rotation
  userMatrix <- structure(c(-0.737, -0.155, 0.568, 0, 0.64, -0.183, 0.94, 0, 0.003, 0.921, 0.363, 0, 0, 0, 0, 1), dim = c(4L, 4L))
  if(save.HTML) userMatrix <- structure(c(-0.746, -0.264, 0.514, 0, 0.625, -0.378, 0.89, 0, 0.002, 0.826, 0.546, 0, 0, 0, 0, 1), dim = c(4L, 4L))
  par3d(userMatrix = userMatrix)
  
  ## Axis labels
  # Expressions cause errors displaying the rglwidget in HTML 20240320
  #xlab <- as.expression(axis.label("T"))
  #ylab <- as.expression(axis.label("O2"))
  # Add axes
  box3d()
  edges <- c("x+", "y+", "z+")
  axes3d(edges)
  mtext3d("Temp (°C)", "x+", line = 4, cex = cex)
  mtext3d("log fO2", "y+", line = 4, cex = cex)
  mtext3d("Zn", "z+", line = 4, cex = cex, at = mean(zlim) + dz)
  mtext3d("(log", "z+", line = 4, cex = cex, at = mean(zlim) - dz)
  mtext3d("ppm)", "z+", line = 4, cex = cex, at = mean(zlim) - 2.5*dz)
  
  # Save as HTML
  if(save.HTML) {
    widget <- rglwidget()
    htmlwidgets::saveWidget(widget, "Zinc_3D.html")
    close3d()
    options(save)
  }
    



• With this Shiny App the user can adjust five variables to see how they affect reaction energy. The base plot was published by Amend et al., 2020. I am responsible only for the app itself and some tweaks to the plot.

  live at shinyapps.io


• The Abundance Animation connects simulated ecological data to environmental change (tap image to open the animation).

  website · code


• Abundance Stacks are used to visualize within-group and between-group variation not only of abundance but also of a chemical metric.

  publication · code


• Deep Time shows parallel trends for evolutionary lineages during the Great Oxidation Event (GOE) and diverging trends during the Neoproterozoic Oxidation Event (NOE). The key design features are the use of a separate panel and color for each lineage and shared colors and scales to indicate features common to both lineages.

  code


• A vertical trend associated with hydration state becomes a sloped trend after virtually Splitting Water into its elements. This visualization is useful for finding the primary chemical variables that influence genome evolution.

  publication · code