import numpy as np
import pandas as pd
try:
ls=[EE_df]
del EE_df
del ls
except NameError:
pass
EE_df=pd.read_csv("./Archived_Data_MNDOT/MPLS_Entry_Exit_Volume_SpecificDates_2.csv")
EE_df.head()
# len(EE_df)
# set(list(EE_df["Sensor"].values))
# EE_df.columns
def modi_Twins_Opener(x):
try:
if (x.split()[1]=="Twins")and(x.split()[2]=="Opener"):
return '2018 Twins Opener'
else:
return x
except:
return x
EE_df["Event Label_2"]=EE_df["Event Label"].apply(modi_Twins_Opener);
EE_df.drop(columns=["Event Label"],inplace=True);
EE_df.rename(columns={"Event Label_2":"Event Label"},inplace=True);
#EE_df.set_index(["Event Label","Sensor"]);
def FromOrTodowntown(x):
def fromto(y):
if y=="from":
return "out"
else:
return "in"
return fromto(x.split()[-2])
def road(x):
a=list(x.split())[0:2]
return ' '.join(a[::-1])
EE_df["Direction(from/to)"]=EE_df["St Label"].apply(FromOrTodowntown);
EE_df["road"]=EE_df["St Label"].apply(road);
EE_df.head();
events=list(EE_df["Event Label"].unique());
events=list(EE_df["Event Label"].unique());
EE_df["sensor_name"]=EE_df["Sensor"]+'('+EE_df["road"]+")";
sensors=list(EE_df["sensor_name"].unique())
EE_df["name_sensor"]=EE_df["road"]+'('+EE_df[ "Sensor"]+")";
sensors=list(EE_df["sensor_name"].unique());
sensors2=list(EE_df["name_sensor"].unique());
len(events);
EE_df["sensor_fromto"]=EE_df["road"]+","+EE_df["Sensor"]+','+EE_df["Direction(from/to)"];
EE_df;
EE_df;
#https://github.com/jupyter-widgets/ipywidgets/issues/1582
from IPython.display import display
import numpy as np
from ipywidgets import *
import matplotlib.pyplot as plt
sensor_out=["S2","S569","S285","S554","S125"]
sensor_in=["S64","S582","S286","S553","S137"]
def plot_flow_event(b):
Num_of_event=events.index(b)
aa=EE_df.groupby(["Event Label"]).get_group(b).sort_values(by="road");
#aa=aa[aa["road"]=="35W"]
aa.set_index("Sensor",inplace=True)
sensors_in=aa.loc[sensor_in,"name_sensor"]
sensors_out=aa.loc[sensor_out,"name_sensor"]
flow_in=aa.loc[sensor_in,"Volume"]
flow_out=aa.loc[sensor_out,"Volume"]
fig, ax = plt.subplots(figsize=(15,5))
opacity=0.8
bar_width = 0.35
index = np.arange(5)
rects1 = plt.bar(index, flow_in, bar_width,
alpha = opacity,
color = 'b',
label = 'To MPLS')
rects2 = plt.bar(index+bar_width, flow_out, bar_width,
alpha = opacity,
color = 'r',
label = 'From MPLS')
# print(index)
# print(sensors_in.values)
# fig, ax = plt.subplots(figsize=(15,5))
# ax.bar(aa["Sensor"],aa["Volume"])
#ax.set_xticklabels(index,tuple(sensors_in.values))
#plt.set_xticklabels(position=[(1.,0),(2.,0)],labels=["A","B"])
#ax.set_xticklabels(index+bar_width,aa["sensor_fromto"],rotation=45,ha="right")
#plt.set_ylabel("vehicle flow")
#plt.set_title(f"Traffic to and from MPLS\n{b}")
#plt.legend()
plt.xlabel('Routes')
plt.ylabel('Volume of Cars')
plt.title(f"Traffic to and from MPLS\n{b}")
plt.xticks(index + bar_width, ('35W_SB(in)', '35W_NB(in)', '394WB(in)', '94WB(in)', '94WB(in)'))
plt.legend()
return None
w1=dict(b=widgets.Dropdown(options=events,value=events[0],description='event',disabled=False))
output = interactive_output(plot_flow_event, w1)
box = VBox([*w1.values(), output])
print("output in an interactive way")
display(box)
#plot_flow_event("October Wed 2018")
def plot_flow_sensor(b):
Num_of_sensor=sensors.index(b)
aa=EE_df.groupby(["sensor_name"]).get_group(b).sort_values(by="Event Label");
#aa=aa[aa["road"]=="35W"]
fig, ax = plt.subplots(figsize=(15,3))
ax.bar(aa["Event Label"],aa["Volume"])
ax.set_xticklabels(aa["Event Label"],rotation=45,ha="right")
ax.set_ylabel("vehicle flow")
ax.set_title(b)
w2=dict(b=widgets.Dropdown(options=sensors,value=sensors[0],description='Sensor',disabled=False))
output = interactive_output(plot_flow_sensor, w2)
box = VBox([*w2.values(), output])
display(box)
def plot_flow_event_all(b,ax):
Num_of_event=events.index(b)
b1,b2=(int(Num_of_event/2),Num_of_event%2)
aa=EE_df.groupby(["Event Label"]).get_group(b).sort_values(by="road");
#aa=aa[aa["road"]=="35W"]
#fig, ax = plt.subplots(figsize=(15,5))
ax[b1,b2].bar(aa["Sensor"],aa["Volume"])
ax[b1,b2].set_xticklabels(aa["sensor_fromto"],rotation=45,ha="right")
if (b2==0):
ax[b1,b2].yaxis.tick_right()
#ax[b1,b2].set_ylabel(b,rotation=0)
box = ax[b1,b2].get_position()
#ax[b1,b2].set_ylabel(b,rotation=45,fontstyle='oblique',position=(box.x0-box.width*0.4,box.y0))
#ax[b1,b2].text(box.x0+box.width*0.4,box.y0,b,rotation=45,fontstyle='oblique')
ax[b1,b2].set_ylabel(b,rotation=0,fontstyle='oblique')
ax[b1,b2].yaxis.label.set_color('red')
else:
ax[b1,b2].set_yticklabels('')
ax[b1,b2].yaxis.set_label_position("right")
ax[b1,b2].set_ylabel(b,rotation=0,fontstyle='oblique')
ax[b1,b2].yaxis.label.set_color('red')
#ax[b1,b2].set_yticklabels(aa['road'],rotation=45,ha="right")
#ax[b1,b2].set_title(b)
#leg[b1,b2]=ax[b1,b2].set_label(b)
return None
fig, ax = plt.subplots(6,2,sharex='all',figsize=(17,8))
# leg=np.zeros((6,2))
for b in events:
plot_flow_event_all(b,ax)
title=fig.suptitle("vechicle flow for 12 sensors",fontsize=22,color="red")
#fig.legend(loc="right")
#fig.legend((leg[0,0],leg[0,1]), ('Line 1', 'Line 2'), 'upper left')
# box = ax[3,1].get_position()
# pos=ax[].set_position([box.x0, box.y0, box.width * 0.8, box.height])
# # Put a legend to the right of the current axis
# leg=ax1.legend(loc='upper left', bbox_to_anchor=(1, 1.0))
Some observations:¶
for events 'October Wed 2018'(10/3/2018 ) and 'Vikings Opener'(9/9/2018), 35 W NB(S64) and 35W SB(S2) are closed. Bothe of these events.
for events 'October Wed 2015', 35W NB(S569),35W SB(S582),394 WB(S285),394 EB(S286) are all closed
- for events October Wed 2016','94 WB(S553),'94 WB(S125) are closed
- We can discern many sensor "pairs", e.g., 35W SB(S2)&35W NB(S64), 35W NB(S569)&35W SB(S582), 394 WB(S285)&394 EB(S286),94 WB(S553)&94 WB(S554), 94 EB(S137)& 94 WB(S125). For each pair, they are on the same highway and near to each other, one in-going and another out-going from the downtown. The flow magnitude in each pair is very close to each other.
##calculate distance between paired sensors
# from math import sin, cos, sqrt, atan2
# R = 6373.0
# lat1 = 52.2296756
# lon1 = 21.0122287
# lat2 = 52.406374
# lon2 = 16.9251681
# dlon = lon2 - lon1
# dlat = lat2 - lat1
# a = (sin(dlat/2))**2 + cos(lat1) * cos(lat2) * (sin(dlon/2))**2
# c = 2 * atan2(sqrt(a), sqrt(1-a))
# distance = R * c
# sensors=list(EE_df["Sensor"].unique())
# sensors
# EE_df.groupby("Sensor")[["Lat","Lng"]].get_group(sensors[0])
sensor_out=["S2","S569","S285","S554","S125"]
sensor_in=["S64","S582","S286","S553","S137"]
for b in events:
aa=EE_df.groupby(["Event Label"]).get_group(b).sort_values(by="road");
#aa["sensor_fromto"].values
# ss=list(aa["Sensor"].values);
# sensors_in=[ss[index_in[i]] for i in range(5)];
# sensors_out=[ss[index_out[i]] for i in range(5)];
# sensors_in
# sensors_out
#aa.loc[sensor_out,"Direction"]
aa.set_index("Sensor",inplace=True)
aa.loc[sensor_in,"name_sensor"]
aa.loc[sensor_in,"Direction"]
#aa.head()
break
index=[4,2,1,3,5,6,9,8,10,7]
35W oSB(S2)&35W NB(S64), 35W NB(S569)&35W SB(S582), 394 WB(S285)&394 EB(S286),94 WB(S553)&94 WB(S554), 94 EB(S137)& 94 WB(S125).
index=[4,2,1,3,5,6,9,8,10,7]
index
index=[ index[i]-1 for i in range(len(index))]
index
index_in=index0[0:9:2]
index_out=index0[1:10:2]
EE_df.groupby("Event Label")["Date"].get_group('Vikings Opener');
index_in
index_out
EE_df