Starting from the example http://pgfplots.net/tikz/examples/contour-and-surface/, I would like to make projection of a surface, defined by a table. The X and Y projections work OK, but the Z projection appears at the top, rather than at the bottom. I guess it is because of the log scale. What do I wrong? The picture below illustrates my problem. (I know, the picture is not attractive and not reasonable, but I wanted to reduce the size for an MWE)
Edition: I also attach an image generated from my full data set. The Y projection is about what I expected. The X projection is funny: just vertical lines, without connection. It looks a bad projection is made: the prohcetd to the Y axis broken lines are projected to the X plane, without any connection. I expected broken lines like the ones in the Y projection. I have no idea what the Z projection is. It looks like it makes a few niveaus around the highest values, and this is all.
I have two differences from the sample: one is the logarithmic scale and the other is the tabular data. I changed the Z scale to linear and setzmin=1e-4, zmax=.1. The result is the third figure. Based on this, I guess that the Y projection is OK, the X projection uses wrong projection direction and Z projection cannot do anything with this wide-range data.
\documentclass[border=10pt]{standalone}
\usepackage{verbatim}
\usepackage{pgfplots}
\pgfplotsset{width=7cm,compat=1.8}
%http://pgfplots.net/tikz/examples/contour-and-surface/
\pgfplotstableset{%
col sep=semicolon,
x index=0,
y index=1,
z index=2,
header=false
}%
\begin{filecontents*}{data/XYZ.csv}
2016; 1; 32.7e-9
2016;10; 1560e-9
2016;20; 9811e-9
2016;25; 5204e-9
2015; 1; 199e-9
2015;10; 438e-9
2015;20; 674e-9
2015;25; 1718e-9
2014; 1; 199e-9
2014;10; 2446e-9
2014;20; 1893e-9
2014;25; 2670e-9
1994; 1; 76957E-9
1994;10;54545455e-9
1994;20; 1390462e-9
1994;25; 1440795e-9
1993; 1; 1167453e-9
1993;10; 7484185e-9
1993;20; 7484185e-9
1993;25; 8876163e-9
\end{filecontents*}
\begin{document}
\begin{tikzpicture}
\pgfplotsset{
every axis/.append style={
scale only axis,
width=\textwidth,
height=\textwidth,
xtick={1995,2000,2005,2010,2015},
ytick={1,10,20,25},
ztick={1e-8,1e-6,1e-4,1e-2,1e-0}
},
/tikz/every picture/.append style={
trim axis left,
trim axis right,
}
}
\begin{axis}
[ %nodes near coords={(\coordindex)},
domain=1980:2020,
domain y=1:40,
footnotesize,
title={Hillside},
/pgf/number format/.cd,
use comma,
1000 sep={},
xlabel=Year,
ylabel=position,
xmin=1980, xmax=2020,% x scale
ymin=1, ymax=40, % y scale
zmin=1e-12, zmax=1, % z scale
xlabel=Year,
zmode = log,
grid=major,
grid style={dotted},
colormap/jet,
zmode=log,
]
\addplot3
[surf]
table {data/XYZ.csv};
% This is the Z (bottom) projection
\addplot3[
contour gnuplot={
% cdata should not be affected by z filter:
output point meta=rawz,
number=15,
labels=false,
},
z filter/.code=\def\pgfmathresult{1E-4},
]
table {data/XYZ.csv};
% This is the X projection
\addplot3[
domain=1980:2020,
domain y=1:30,
% we want 1d (!) individually colored mesh segments:
mesh, patch type=line,
x filter/.code=\def\pgfmathresult{1980},
]
table {data/XYZ.csv};
%% This is the Y projection
\addplot3[
% we want 1d (!) individually colored mesh segments:
mesh, patch type=line,
y filter/.code=\def\pgfmathresult{40},
]
table {data/XYZ.csv};
\end{axis}
\end{tikzpicture}
\end{document}
XYZ.csvfile? – Alenanno Jul 31 '16 at 09:54x exprand the like. I just found that usingx filteris the documented way. – Henri Menke Aug 01 '16 at 08:27