kyle.belanger/learning_Julia
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions

update graph

Browse source
This commit is contained in:
Kyle Belanger 2024-08-13 14:28:21 -04:00
parent 819543626c
commit 2ac17a08e2
1 changed files with 25 additions and 7 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

30
LearnJulia/src/amtrak.jl
View file

@ -5,8 +5,11 @@ using DataFrames
using DataFramesMeta
using Dates
using Statistics
using AlgebraOfGraphics
using CairoMakie
url = "https://juckins.net/amtrak_status/archive/html/history.php?train_num=97&station=&date_start=07%2F01%2F2024&date_end=07%2F31%2F2024&df1=1&df2=1&df3=1&df4=1&df5=1&df6=1&df7=1&sort=schDp&sort_dir=DESC&co=gt&limit_mins=&dfon=1"
url = "https://juckins.net/amtrak_status/archive/html/history.php?train_num=97%2C98&station=&date_start=07%2F01%2F2024&date_end=07%2F31%2F2024"
resp = HTTP.get(url)
@ -16,7 +19,7 @@ s = sel"tr"
rows = eachmatch(s, page.root)
# create empty DataFrame and then populate it with the table from website
df = DataFrame(orgin_date = [], station = [], sch_dp = [], act_dp = String[], comments = [], s_disrupt = [], cancellations = [])
df = DataFrame(train = String[], orgin_date = [], station = String[], sch_dp = [], act_dp = String[], comments = [], s_disrupt = [], cancellations = [])
for i in rows
text = eachmatch(Selector("td"), i)
@ -47,20 +50,35 @@ end
mod_df = @chain df begin
@rsubset :act_dp != "" && :s_disrupt != "SD"
@select :station :comments
@select :train :station :comments
#can't perform match if there is nothing there
@rtransform :delay = if occursin(r"Dp:", :comments) match(r"Dp:.*", :comments).match else "" end
@rtransform :min = if occursin(r"min", :delay) match(r"[0-9]* min", :delay).match |>
x -> parse(Int,match(r"[0-9]*", x).match) else Int(0) end
@rtransform :hour = if occursin(r"hr", :delay) match(r"[1-9]* hr", :delay).match |>
x -> parse(Int,match(r"[1-9]*", x).match) |> x -> x*60 else Int(0) end
@rtransform :total_delay_mins = :min + :hour
@rtransform :total_delay_mins = :min + :hour |> x -> ifelse(occursin(r"late", :delay), x, x *-1)
end
gd = @by mod_df :station begin
:mean = Statistics.mean(:total_delay_mins)
gd = @chain mod_df begin
@by _ [:train,:station] begin
:mean = Float32[Statistics.mean(:total_delay_mins)]
:median = Statistics.median(:total_delay_mins)
:max = maximum(:total_delay_mins)
:min = minimum(:total_delay_mins)
end
@orderby :station :train
@groupby :station
@transform :diff = [missing; diff(:mean)]
end
# plot(bar(gd.station, gd.mean), xticks = (1:length(gd.station), gd.station), xrotation = 90, legend = false, title = "Mean Delay by Station", xlabel = "Station", ylabel = "Delay (mins)")
axis = (width = 750, height = 750, title = "Mean Delay by Station", xlabel = "Station", ylabel = "Delay (mins)",xticklabelrotation = 45)
mean_delay = data(gd) * mapping(:station, :mean, color = :train => "Train", dodge = :train) * visual(BarPlot)
draw(mean_delay; axis = axis)