I need to plot something that looks like a gant chart, but are actual data ranges. These ranges need to be on a log scale. Graphics does not offer a Scalingfunction. A LogPlot with ErrorBars does not work either and would not look ideal. Is it possible to draw rectangles with one axis being log scale?
E.g.: Min and Max values of the rectangles are the known data.
SeedRandom[1]
dates = Sort /@ RandomInteger[{10, 1000}, {5, 2}];
data = MapIndexed[Thread@{#, #2[[1]]} &, dates];
labels = CharacterRange["A", "E"];
data2 = MapIndexed[Labeled[{#, #2[[1]]}, Style[labels[[#2[[1]]]], White,
FontSize -> Scaled[.03]], Center] &, GeometricMean /@ dates];
ListLogLinearPlotShow[ListLogLinearPlot[data,
BaseStyle -> Directive[AbsoluteThickness[36], CapForm["Butt"]],
Joined -> True, AspectRatio -> 1/2, ImageSize -> Large,
PlotRange -> {0, 6}, Frame -> True,
FrameTicks -> {{None, None}, {{50, 100, 200, 500, 1000}, Automatic}}],
ListLogLinearPlot[data2, PlotMarkers -> ""]]
ScalingFunctionsListLinePlot[data, BaseStyle -> Directive[AbsoluteThickness[36], CapForm["Butt"]],
ScalingFunctions -> {"Log", "Linear"}, AspectRatio -> 1/2,
ImageSize -> Large, PlotRange -> {0, 6}, Frame -> True,
FrameTicks -> {{None, None}, {{50, 100, 200, 500, 1000}, Automatic}},
Epilog -> MapIndexed[Text[Style[labels[[#2[[1]]]], White,
FontSize -> Scaled[.03]], {#, #2[[1]]}, Center] &, Log[GeometricMean /@ dates]]]
TimeLinePlotTimelinePlot[{Labeled[#, Style[#2, White, FontSize -> Scaled[.03]], Center]} & @@@
Transpose[{Interval /@ (Map[N@*Log, dates, {-1}]), labels}],
PlotStyle -> Directive[AbsoluteThickness[40], CapForm["Butt"]],
Spacings -> .1, PlotMarkers -> "", Frame -> True,
AspectRatio -> 1/2, PerformanceGoal -> "Speed",
FrameTicks -> {{Automatic, Automatic},
{Transpose[{Log@#, #} &@{50, 100, 200, 500}], None}}]
BarChartBarChart[{Style[#[[1]], Directive[EdgeForm[], White]],
Labeled[Style[#[[2]], Directive[EdgeForm[], #2]],
Style[#3, White, FontSize -> Scaled[.03]], Center]} & @@@
Transpose[{dates, ColorData[97] /@ Range[Length@dates], labels}],
BarOrigin -> Left, BarSpacing -> {0, .25}, ChartLayout -> "Stacked",
ScalingFunctions -> "Log", Axes -> False,
PerformanceGoal -> "Speed", Frame -> True, ImageSize -> Large,
PlotRangePadding -> .5, AspectRatio -> 1/2]
Thanks: @David G. Stork for the GeometricMean idea to center the labels.
See also: Poets of the 19th century
Mean is simply incorrect. Try GeometricMean.
$\endgroup$
Commented
Mar 15, 2019 at 21:24
LogLinearPlot[1,
{t, .1, 100},
PlotStyle -> White,
Epilog -> {Rectangle[{Log[1], .5}, {Log[50], 1}],
Red, Rectangle[{Log[30], 1.5}, {Log[80], 2}],
Text[Style["a", White, 18], {Log[7], .75}],
Text[Style["b", White, 18], {Log[53], 1.75}]}]
Rectangle[]. Thank you!
$\endgroup$
A = 10^5; B = 5;
rectangle[start_, length_, number_] :=
RegionPlot[Log10[start] <= x <= Log10[start + length] &&
number <= y <= number + .75, {x, 0, Log10[A]}, {y, 0, B},
PlotStyle -> LightBlue, BoundaryStyle -> Blue];
Show[
rectangle[10, 300, 1], rectangle[100, 3000, 2], rectangle[1000, 30000, 3],
FrameTicks -> {Transpose@{Range[0,5], 10^Range[0,5]}, Automatic, None, None},
AspectRatio -> 1/2
]
FrameTicks[]. You are right, not a problem.
$\endgroup$