shrdlu
March 28, 2024, 6:23pm
1
Hi,
I’m trying the AlgebraOfGraphics page demo for heatmap-style visualization (Visual · Algebra of Graphics ) using my own local data.
I have no issues importing the external .tsv matrix file, but the process throws an error when it comes to drawing the actual plot. Here’s the code I’m trying to get to work.
using AlgebraOfGraphics, DataFrames, CairoMakie, CSV
set_aog_theme!()
heat_matrix = CSV.read("localdata.tsv", DataFrame)
plt = data(heat_matrix) * mapping(:x, :y) * AlgebraOfGraphics.density(npoints=50)
draw(plt * visual(colormap=:viridis))
The error message is:
julia> draw(plt * visual(colormap=:viridis))
ERROR: ArgumentError: column name :x not found in the data frame
And the input data is roughly in form below, though the number of entries is about 876 by 876
entry_1 entry_2 entry_3 entry_4 entry_5 entry_6 entry_7 entry_8 entry_9 entry_10 entry_11 entry_12 entry_13 entry_14 entry_15
100.0 99.389389 80.807358 81.091492 80.045761 80.2612 80.591591 80.326904 80.517456 80.73204 80.916733 81.039085 80.428047 81.411896 82.012657
99.389389 100.0 80.769554 81.563904 80.253723 81.239258 80.34507 80.411316 80.983093 80.776138 80.54747 80.830902 80.215294 81.910057 82.269379
80.807358 80.769554 100.0 98.784981 87.967583 88.039368 88.1138 88.004013 84.194122 84.372253 84.497612 85.310501 82.93158 82.449257 82.363342
81.091492 81.563904 98.784981 100.0 88.240082 88.462433 87.531174 87.579102 83.479767 83.909073 84.052917 84.921959 83.114807 82.751495 82.161125
80.045761 80.253723 87.967583 88.240082 100.0 98.189163 88.235207 89.16333 84.933044 84.678314 86.075043 86.235718 83.413078 82.43309 80.658096
80.2612 81.239258 88.039368 88.462433 98.189163 100.0 88.264221 88.903702 85.049088 85.365807 86.156403 86.378174 83.042229 82.283073 80.690521
80.591591 80.34507 88.1138 87.531174 88.235207 88.264221 100.0 89.984344 84.083481 83.85466 87.144073 85.681107 83.881477 81.86351 82.695709
80.326904 80.411316 88.004013 87.579102 89.16333 88.903702 89.984344 100.0 85.202805 84.645584 84.900604 85.778015 85.173973 81.289238 82.033951
80.517456 80.983093 84.194122 83.479767 84.933044 85.049088 84.083481 85.202805 100.0 99.035522 87.151443 88.334236 84.80954 86.120651 83.791351
80.73204 80.776138 84.372253 83.909073 84.678314 85.365807 83.85466 84.645584 99.035522 100.0 87.532318 88.439926 84.654488 85.649384 83.718521
80.916733 80.54747 84.497612 84.052917 86.075043 86.156403 87.144073 84.900604 87.151443 87.532318 100.0 92.740463 85.269989 82.425095 83.434967
81.039085 80.830902 85.310501 84.921959 86.235718 86.378174 85.681107 85.778015 88.334236 88.439926 92.740463 100.0 85.421165 82.606354 83.663094
80.428047 80.215294 82.93158 83.114807 83.413078 83.042229 83.881477 85.173973 84.80954 84.654488 85.269989 85.421165 100.0 84.637268 84.210716
81.411896 81.910057 82.449257 82.751495 82.43309 82.283073 81.86351 81.289238 86.120651 85.649384 82.425095 82.606354 84.637268 100.0 84.265274
82.012657 82.269379 82.363342 82.161125 80.658096 80.690521 82.695709 82.033951 83.791351 83.718521 83.434967 83.663094 84.210716 84.265274 100.0
The dataframe clearly shows the first line as the header for each column - am pretty stumped at the moment. What could I be doing wrong?
Thank you!
plt = data(heat_matrix) * mapping(:entry_1) * AlgebraOfGraphics.density(npoints=50)
Your column names are entry_j. There are, as the error message says, no columns named :x or :y.
The mapping for density takes a single argument. See
https://tutorials.pumas.ai/html/PlottingInJulia/03-AoG-Stats.html#density
1 Like
shrdlu
March 29, 2024, 3:41pm
3
Thank you for responding - do you know if there’s a way to simply plot out, say, 800x800 matrix table of number on x-y axis heatmap in AoG, with values corresponding to color range? I can’t seem to find an actual example to study…
I get a feeling the makie-AoG idea of heatmap in the linked page on the opening post isn’t meant to address what I’m trying to do, since it’s making a default assumption of linear corresponding 2 column data input. And the type of matrix file I’m trying to plot as heatmap can be tens of thousands of lines across x-y axis.
Just to clarify with some real world examples - attaching something I threw together in about a minute in gnuplot using a similar dataset as in the post:
And here’s a Plots.jl output with GR backend, using the same data (not cropped) I posted here:
Any lead would be helpful - so far all I’ve been finding are witty one-liners without any explanations of intent or what’s being expected for the package, IMHO.
jar1
March 29, 2024, 6:46pm
4
data = [
100.0 99.389389 80.807358 81.091492 80.045761 80.2612 80.591591 80.326904 80.517456 80.73204 80.916733 81.039085 80.428047 81.411896 82.012657
99.389389 100.0 80.769554 81.563904 80.253723 81.239258 80.34507 80.411316 80.983093 80.776138 80.54747 80.830902 80.215294 81.910057 82.269379
80.807358 80.769554 100.0 98.784981 87.967583 88.039368 88.1138 88.004013 84.194122 84.372253 84.497612 85.310501 82.93158 82.449257 82.363342
81.091492 81.563904 98.784981 100.0 88.240082 88.462433 87.531174 87.579102 83.479767 83.909073 84.052917 84.921959 83.114807 82.751495 82.161125
80.045761 80.253723 87.967583 88.240082 100.0 98.189163 88.235207 89.16333 84.933044 84.678314 86.075043 86.235718 83.413078 82.43309 80.658096
80.2612 81.239258 88.039368 88.462433 98.189163 100.0 88.264221 88.903702 85.049088 85.365807 86.156403 86.378174 83.042229 82.283073 80.690521
80.591591 80.34507 88.1138 87.531174 88.235207 88.264221 100.0 89.984344 84.083481 83.85466 87.144073 85.681107 83.881477 81.86351 82.695709
80.326904 80.411316 88.004013 87.579102 89.16333 88.903702 89.984344 100.0 85.202805 84.645584 84.900604 85.778015 85.173973 81.289238 82.033951
80.517456 80.983093 84.194122 83.479767 84.933044 85.049088 84.083481 85.202805 100.0 99.035522 87.151443 88.334236 84.80954 86.120651 83.791351
80.73204 80.776138 84.372253 83.909073 84.678314 85.365807 83.85466 84.645584 99.035522 100.0 87.532318 88.439926 84.654488 85.649384 83.718521
80.916733 80.54747 84.497612 84.052917 86.075043 86.156403 87.144073 84.900604 87.151443 87.532318 100.0 92.740463 85.269989 82.425095 83.434967
81.039085 80.830902 85.310501 84.921959 86.235718 86.378174 85.681107 85.778015 88.334236 88.439926 92.740463 100.0 85.421165 82.606354 83.663094
80.428047 80.215294 82.93158 83.114807 83.413078 83.042229 83.881477 85.173973 84.80954 84.654488 85.269989 85.421165 100.0 84.637268 84.210716
81.411896 81.910057 82.449257 82.751495 82.43309 82.283073 81.86351 81.289238 86.120651 85.649384 82.425095 82.606354 84.637268 100.0 84.265274
82.012657 82.269379 82.363342 82.161125 80.658096 80.690521 82.695709 82.033951 83.791351 83.718521 83.434967 83.663094 84.210716 84.265274 100.0
]
using CairoMakie
heatmap(data)
3 Likes
I think you’re correct in that Heatmap for AoG is more for creating 2D histrograms/density plots. I would do what @jar1 suggests but if you really want to use AoG you could the geometry example :
using AlgebraOfGraphics, CairoMakie
using GeometryBasics
entries = [100.0 99.389389 80.807358 81.091492 80.045761 80.2612 80.591591 80.326904 80.517456 80.73204 80.916733 81.039085 80.428047 81.411896 82.012657
99.389389 100.0 80.769554 81.563904 80.253723 81.239258 80.34507 80.411316 80.983093 80.776138 80.54747 80.830902 80.215294 81.910057 82.269379
80.807358 80.769554 100.0 98.784981 87.967583 88.039368 88.1138 88.004013 84.194122 84.372253 84.497612 85.310501 82.93158 82.449257 82.363342
81.091492 81.563904 98.784981 100.0 88.240082 88.462433 87.531174 87.579102 83.479767 83.909073 84.052917 84.921959 83.114807 82.751495 82.161125
80.045761 80.253723 87.967583 88.240082 100.0 98.189163 88.235207 89.16333 84.933044 84.678314 86.075043 86.235718 83.413078 82.43309 80.658096
80.2612 81.239258 88.039368 88.462433 98.189163 100.0 88.264221 88.903702 85.049088 85.365807 86.156403 86.378174 83.042229 82.283073 80.690521
80.591591 80.34507 88.1138 87.531174 88.235207 88.264221 100.0 89.984344 84.083481 83.85466 87.144073 85.681107 83.881477 81.86351 82.695709
80.326904 80.411316 88.004013 87.579102 89.16333 88.903702 89.984344 100.0 85.202805 84.645584 84.900604 85.778015 85.173973 81.289238 82.033951
80.517456 80.983093 84.194122 83.479767 84.933044 85.049088 84.083481 85.202805 100.0 99.035522 87.151443 88.334236 84.80954 86.120651 83.791351
80.73204 80.776138 84.372253 83.909073 84.678314 85.365807 83.85466 84.645584 99.035522 100.0 87.532318 88.439926 84.654488 85.649384 83.718521
80.916733 80.54747 84.497612 84.052917 86.075043 86.156403 87.144073 84.900604 87.151443 87.532318 100.0 92.740463 85.269989 82.425095 83.434967
81.039085 80.830902 85.310501 84.921959 86.235718 86.378174 85.681107 85.778015 88.334236 88.439926 92.740463 100.0 85.421165 82.606354 83.663094
80.428047 80.215294 82.93158 83.114807 83.413078 83.042229 83.881477 85.173973 84.80954 84.654488 85.269989 85.421165 100.0 84.637268 84.210716
81.411896 81.910057 82.449257 82.751495 82.43309 82.283073 81.86351 81.289238 86.120651 85.649384 82.425095 82.606354 84.637268 100.0 84.265274
82.012657 82.269379 82.363342 82.161125 80.658096 80.690521 82.695709 82.033951 83.791351 83.718521 83.434967 83.663094 84.210716 84.265274 100.0]
geometry = [Rect(Vec(i, j), Vec(1, 1)) for i in 1:15 for j in 1:15]
group = vec(entries)
df = (; geometry, group)
plt = data(df) * visual(Poly) * mapping(:geometry, color = :group)
fg = draw(plt; axis=(aspect=1,))
Which produces:
1 Like
jules
March 30, 2024, 9:14am
6
AoG can do heatmaps, the thing is just that it expects long format, so you cannot have a matrix-like dataframe but need to transform it into something that has an x, y and value column. In your case that could look like this (I’m not using AoG here but this shows off the heatmap method of Makie that you would use):
using DataFrames
using CSV
using CairoMakie
using Chain
data = """
entry_1 entry_2 entry_3 entry_4 entry_5 entry_6 entry_7 entry_8 entry_9 entry_10 entry_11 entry_12 entry_13 entry_14 entry_15
100.0 99.389389 80.807358 81.091492 80.045761 80.2612 80.591591 80.326904 80.517456 80.73204 80.916733 81.039085 80.428047 81.411896 82.012657
99.389389 100.0 80.769554 81.563904 80.253723 81.239258 80.34507 80.411316 80.983093 80.776138 80.54747 80.830902 80.215294 81.910057 82.269379
80.807358 80.769554 100.0 98.784981 87.967583 88.039368 88.1138 88.004013 84.194122 84.372253 84.497612 85.310501 82.93158 82.449257 82.363342
81.091492 81.563904 98.784981 100.0 88.240082 88.462433 87.531174 87.579102 83.479767 83.909073 84.052917 84.921959 83.114807 82.751495 82.161125
80.045761 80.253723 87.967583 88.240082 100.0 98.189163 88.235207 89.16333 84.933044 84.678314 86.075043 86.235718 83.413078 82.43309 80.658096
80.2612 81.239258 88.039368 88.462433 98.189163 100.0 88.264221 88.903702 85.049088 85.365807 86.156403 86.378174 83.042229 82.283073 80.690521
80.591591 80.34507 88.1138 87.531174 88.235207 88.264221 100.0 89.984344 84.083481 83.85466 87.144073 85.681107 83.881477 81.86351 82.695709
80.326904 80.411316 88.004013 87.579102 89.16333 88.903702 89.984344 100.0 85.202805 84.645584 84.900604 85.778015 85.173973 81.289238 82.033951
80.517456 80.983093 84.194122 83.479767 84.933044 85.049088 84.083481 85.202805 100.0 99.035522 87.151443 88.334236 84.80954 86.120651 83.791351
80.73204 80.776138 84.372253 83.909073 84.678314 85.365807 83.85466 84.645584 99.035522 100.0 87.532318 88.439926 84.654488 85.649384 83.718521
80.916733 80.54747 84.497612 84.052917 86.075043 86.156403 87.144073 84.900604 87.151443 87.532318 100.0 92.740463 85.269989 82.425095 83.434967
81.039085 80.830902 85.310501 84.921959 86.235718 86.378174 85.681107 85.778015 88.334236 88.439926 92.740463 100.0 85.421165 82.606354 83.663094
80.428047 80.215294 82.93158 83.114807 83.413078 83.042229 83.881477 85.173973 84.80954 84.654488 85.269989 85.421165 100.0 84.637268 84.210716
81.411896 81.910057 82.449257 82.751495 82.43309 82.283073 81.86351 81.289238 86.120651 85.649384 82.425095 82.606354 84.637268 100.0 84.265274
82.012657 82.269379 82.363342 82.161125 80.658096 80.690521 82.695709 82.033951 83.791351 83.718521 83.434967 83.663094 84.210716 84.265274 100.0
"""
@chain begin
data
IOBuffer
CSV.read(DataFrame; delim = '\t')
transform(eachindex => :x)
stack(Not(:x))
transform(:variable => ByRow(x -> parse(Int, split(x, "_")[2])) => :y)
@aside display(first(_, 10))
heatmap(_.x, _.y, _.value; axis = (; aspect = DataAspect()))
end
This prints out
10×4 DataFrame
Row │ x variable value y
│ Int64 String Float64 Int64
─────┼──────────────────────────────────
1 │ 1 entry_1 100.0 1
2 │ 2 entry_1 99.3894 1
3 │ 3 entry_1 80.8074 1
4 │ 4 entry_1 81.0915 1
5 │ 5 entry_1 80.0458 1
6 │ 6 entry_1 80.2612 1
7 │ 7 entry_1 80.5916 1
8 │ 8 entry_1 80.3269 1
9 │ 9 entry_1 80.5175 1
10 │ 10 entry_1 80.732 1
and the plot looks like
3 Likes
