
import pandas as pd
from scipy import stats
import numpy as np
import matplotlib.pyplot as plt
import seaborn as sns
import folium 
from folium.plugins import TimestampedGeoJson
from sklearn.model_selection import train_test_split
from sklearn import linear_model
from sklearn import preprocessing, svm
from sklearn.linear_model import LogisticRegression
from sklearn import metrics
from sklearn.ensemble import RandomForestClassifier


df = pd.read_csv("accidentData.csv")

df


df = df.dropna()
df


count = {}
for x in range(943318):
    count[df.iat[x,2][0:4]] = count.get(df.iat[x,2][0:4],0)+1

count_sorted = dict(sorted(count.items()))
count_sorted

{'2016': 693,
 '2017': 1590,
 '2018': 2225,
 '2019': 49694,
 '2020': 219585,
 '2021': 669531}


plt.figure(figsize=(10,5))
plt.bar(list(count_sorted.keys()),list(count_sorted.values()), color= ['#3399FF'])
plt.xlabel("Year")
plt.ylabel("No. of Accidents")
plt.title("No. of Accidents per Year")
plt.show()


toBeRemoved = []
for x in range(943318):
    if df.iat[x,2][0:4] != "2020":
        toBeRemoved.append(df.index[x])
df = df.drop(toBeRemoved)
df


df.columns

Index(['ID', 'Severity', 'Start_Time', 'End_Time', 'Start_Lat', 'Start_Lng',
       'End_Lat', 'End_Lng', 'Distance(mi)', 'Description', 'Number', 'Street',
       'Side', 'City', 'County', 'State', 'Zipcode', 'Country', 'Timezone',
       'Airport_Code', 'Weather_Timestamp', 'Temperature(F)', 'Wind_Chill(F)',
       'Humidity(%)', 'Pressure(in)', 'Visibility(mi)', 'Wind_Direction',
       'Wind_Speed(mph)', 'Precipitation(in)', 'Weather_Condition', 'Amenity',
       'Bump', 'Crossing', 'Give_Way', 'Junction', 'No_Exit', 'Railway',
       'Roundabout', 'Station', 'Stop', 'Traffic_Calming', 'Traffic_Signal',
       'Turning_Loop', 'Sunrise_Sunset', 'Civil_Twilight', 'Nautical_Twilight',
       'Astronomical_Twilight'],
      dtype='object')


df = df.drop(['End_Lat', 'End_Lng','Distance(mi)', 'Number', 'Street',
       'Side', 'City', 'County','Zipcode', 'Country', 'Timezone',
       'Airport_Code', 'Weather_Timestamp','Civil_Twilight', 'Nautical_Twilight',
       'Astronomical_Twilight'], axis=1)
df


plt.figure(figsize=(10,5))
sev = sns.countplot(data= df, x= 'Severity', 
                    palette= ['#ffe330', '#ffb72c', '#ff8628', '#ff4224'])
sev.set(ylabel= "No. of Accidents", title= "Accidents per Severity")
plt.show()


plt.figure(figsize=(13,5))
sns.violinplot(x ="Severity", y = "Temperature(F)", data = df)
plt.title("Temperature(F) vs Accident Serverity:")
plt.show()


plt.figure(figsize=(11,4))
sns.violinplot(x ="Severity", y = "Visibility(mi)", data = df)
plt.title("Visibility(mi) vs Accident Severity:")
plt.show()


# Code to remove outliers more than four std devs from the mean
mean_vis = df['Visibility(mi)'].mean()
std_vis = df['Visibility(mi)'].std()
clean_vis = df[(df['Visibility(mi)'] <= mean_vis + (4*std_vis))]
clean_vis = clean_vis[(clean_vis['Visibility(mi)'] >= mean_vis - (4 * std_vis))]

# Plot cleaned data
plt.figure(figsize=(11,4))
sns.violinplot(x ="Severity", y = "Visibility(mi)", data = clean_vis)
plt.title("Visibility(mi) vs Accident Severity:")
plt.show()


plt.figure(figsize=(13,6))
sns.violinplot(x ="Severity", y = "Wind_Speed(mph)", data = df)
plt.title("Wind_Speed(mph) vs Accident Serverity:")
plt.show()


# Code to remove outliers more than 4 standard devs. from mean
mean_wind = df['Wind_Speed(mph)'].mean()
std_wind = df['Wind_Speed(mph)'].std()
clean_wind = df[(df['Wind_Speed(mph)'] <= mean_wind + (4*std_wind))]
clean_wind = clean_wind[(clean_wind['Wind_Speed(mph)'] >= mean_wind - (4 * std_wind))]

# Plot new data
plt.figure(figsize=(13,6))
sns.violinplot(x ="Severity", y = "Wind_Speed(mph)", data = clean_wind)
plt.title("Wind Speed(mph) vs Accident Serverity:")
plt.ylabel("Wind Speed (mph)")
plt.show()


plt.figure(figsize = (30, 10))
sns.set(font_scale = 1.1)
accidentplot = sns.countplot(data= df, x= 'Weather_Condition', 
                             order= df['Weather_Condition'].value_counts().index)
accidentplot.set_xticklabels(accidentplot.get_xticklabels(), rotation=40, ha="right")
accidentplot.set_title('Weather Conditions During Accidents')
accidentplot.set(xlabel= 'Weather Condition', ylabel= 'Number of Accidents')
plt.show()


count = {}
for x in range(219585):
    count[df.iat[x,12]] = count.get(df.iat[x,12],0)+1
# Copied directly from count variable, showing largest categories
updatedCount = {"Cloudy":37293, "Fair":104828, "Partly Cloudy":18402, 
                "Fog":4321, "Light Rain":9825, "Mostly Cloudy":29139}

topsix = pd.DataFrame(list(updatedCount.items()), columns=['Weather', 'Count'])
topsix = topsix.sort_values('Count', ascending=False)

plt.figure(figsize=(12,5))
six_bar = sns.barplot(x= 'Weather', y= 'Count', data= topsix, 
                      palette= ['#ff0000', '#d7002e', '#af0047', '#86005c', 
                                '#58006f', '#000080'])
six_bar.set_xticklabels(six_bar.get_xticklabels(), rotation=40, ha="right")
six_bar.set_title("Top Weather Conditions During Accidents")
six_bar.set(xlabel= 'Weather Condition', ylabel= "Number of Accidents")
six_bar.tick_params(axis='both', which='major', labelsize=10)


# Count number of accidents corresponding to day or night
count = {}
for x in range(219585):
    count[df.iat[x,30]] = count.get(df.iat[x,30],0)+1

plt.figure(figsize=(8,3))
plt.bar(list(count.keys()),list(count.values()), color  = ['#e28743', "#1e81b0"])
plt.xlabel("Time of Day")
plt.ylabel("No. of Accidents")
plt.title("No. of Accidents per Time of Day")
plt.show()


# Set cols to datetime
df["Start_Time"] =  pd.to_datetime(df["Start_Time"])
df["End_Time"] =  pd.to_datetime(df["End_Time"])


# Count accidents per hour
count = {}
for x in range(219585):
    count[df.iat[x,2].hour] = count.get(df.iat[x,2].hour,0)+1

sorted_count = dict(sorted(count.items()))
tod = pd.DataFrame.from_dict(sorted_count, orient='index', columns=['Accidents'])


plt.figure(figsize=(15,6))
tod_plot = sns.barplot(x= tod.index.values, y= 'Accidents', data= tod, 
                       palette= ['#ffa500', '#ff9902', '#fe8d09', '#fc8113', '#f9761b', 
                                 '#f46b23', '#ef602a', '#e85531', '#e14b37', '#da413d', 
                                 '#d13743', '#c82e4a', '#bf2550', '#b41b56', '#a9125c', 
                                 '#9e0761', '#910067', '#83006d', '#730072', '#640077', 
                                 '#53007b', '#40007e', '#2a007f', '#000080'])
tod_plot.set(title= "Car Accidents per Hour in 2021", ylabel= "Accidents", xlabel= "Hour")
plt.show()


# Add cols. for start and end hour for each var. This will be useful later.
df['Start_Hour'] = df['Start_Time'].apply(lambda x: x.hour)
df['End_Hour'] = df['End_Time'].apply(lambda x: x.hour)
df = df.dropna()
df.Start_Hour.astype(int)
df.End_Hour.astype(int)

plt.figure(figsize=(15,6))
tod_sev_plot = sns.countplot(x= 'Start_Hour', hue= 'Severity', data= df, 
                                palette= ['#b9f174', '#ffb42a', '#ff8424', '#ff4224'])
tod_sev_plot.set(xlabel= "Start Hour", ylabel= "No. of Accidents", 
                        title= "Severity of Accidents per Hour")
plt.show()


# Find severe (3 or 4) level accidents per hour
severe_arr = []
for i in range(24):
    temp = df[df['Start_Hour'] == i]
    total = len(temp.index)
    sev = len(temp[temp['Severity'] >= 3])
    severe_arr.append([sev, total, sev/total])
severe_df= pd.DataFrame(data= severe_arr, columns= ['Severe', 'Total', 'Proportion'])

# Plot data
plt.figure(figsize=(15,7))
x = sns.lineplot(data= severe_df, x= severe_df.index.values, y= 'Proportion')
x.set(xlabel= 'Hour', title= 'Proportion of Severe Auto Accidents')
plt.show()


temp = df[df['Severity'] > 2]

accident_map = folium.Map(location=[37.0902, -95.7129], zoom_start=4.45, tiles='Stamen Terrain', prefer_canvas = True,zoom_control=False,
               scrollWheelZoom=False, dragging=False)
#Used to set color of severity of accident on map
def color_manager(severity):
    if severity  == 1:
        return 'lightgreen'
    if severity == 2:
        return 'yellow'
    if severity == 3:
        return 'orange'
    else:
        return 'red'
        
for x in range(18579):
    folium.CircleMarker(
        location=[temp.iat[x, 4],temp.iat[x, 5]],
        radius=2,
        fill = True,
        fill_opacity = 0.3, # Setting the inner circle opacity
        color = color_manager(temp.iat[x,1]),
        opacity = 0.4,
        bubling_mouse_events = True,
        popup= [temp.iat[x,6]] # Setting up an info pop up with Event type as it's info
    ).add_to(accident_map)
accident_map


def linregML(X,Y, dep, ind, title):
    # Split data into testing and training 
    X_train, X_test, y_train, y_test = train_test_split(X, Y, test_size = 0.25)

    # Create regression
    regr = linear_model.LinearRegression()

    # Fit data by training model
    regr.fit(X_train, y_train)

    print('Regression Score: %.8f' % regr.score(X_test, y_test))

    # Make predictions using the test group
    y_pred = regr.predict(X_test)
    
    plt.figure(figsize= (15,7))
    plt.scatter(X_test, y_test, color ='b', s= 0.5)
    plt.plot(X_test, y_pred, color ='k')
    plt.ylim((0,500))

    plt.title(title)
    plt.xlabel(dep)
    plt.ylabel(ind)

    plt.show()


def RANSAClinregML(X,Y, dep, ind, title):
    ransac = linear_model.RANSACRegressor()
    ransac.fit(X, Y)
    inlier_mask = ransac.inlier_mask_
    outlier_mask = np.logical_not(inlier_mask)

    Xin = X[inlier_mask]
    Yin = Y[inlier_mask]

    # split the data into training/testing sets
    X_train, X_test, y_train, y_test = train_test_split(Xin, Yin, test_size=0.25)
    regr = linear_model.LinearRegression()
    regr.fit(X_train, y_train)

    print('Regression Score: %.8f' % regr.score(X_test, y_test))

    plt.figure(figsize= (15,7))
    y_pred = regr.predict(X_test)
    plt.scatter(X_test, y_test, color ='b', s= 0.5)
    plt.plot(X_test, y_pred, color ='k')

    plt.title(title)
    plt.xlabel(dep)
    plt.ylabel(ind)

    plt.show()


def logML(X,Y):
    X_train, X_test, y_train, y_test = train_test_split(X, Y, test_size=0.25, random_state=16) #random state to sample the data randomly
    logreg = LogisticRegression(random_state=16) #same here

    logreg.fit(X_train, y_train)
    y_pred = logreg.predict(X_test)

    #could add a solver here if the penalty is Type 2 -> lbfgs & saga/sag -> type 1 error. 
    # depends on the matrix down below

    #Can also add a ROC curve to see the rate of false positives and false negatives

    #creating a heatmap confusion matrix to visually represent the logistic regression

    heat_matrix = metrics.confusion_matrix(y_test, y_pred)

    fig, ax = plt.subplots()

    sns.heatmap(pd.DataFrame(heat_matrix), annot=True, cmap="YlGnBu" ,fmt='g')
    ax.xaxis.set_label_position("top")
    plt.tight_layout()
    plt.title('Confusion matrix', y=1.1)
    plt.ylabel('Actual label')
    plt.xlabel('Predicted label')

    plt.plot()

    #printint the metrics associated to heatmaps for further analysis
    print('Classification Accuracy: %.2f', (heat_matrix[0,0] + heat_matrix[0,1])/(heat_matrix[0,0] +heat_matrix[0,1] +heat_matrix[1,0] +heat_matrix[1,1]))

    precision = heat_matrix[1,1]/(heat_matrix[1,1] + heat_matrix[0,1])
    recall = heat_matrix[1,1]/(heat_matrix[1,1] + heat_matrix[1,0])
    print('Precision: %.2f', precision)
    print('Recall: %.2f', recall)
    print('F1 Score: %.2f', (2*precision*recall)/(precision + recall))


# Using duration instead of severity as mentioned in the OLS explanation
df['Duration'] = df.apply(
    lambda x: pd.Timedelta(x['End_Time']-x['Start_Time']).total_seconds() / 60, axis=1)


df['Duration'] = pd.DataFrame(df['Duration'])
df = df[df['Duration'] <= 2880]
df = df[df['Duration'] > 10]

# Now, apply to models
x = df[['Humidity(%)']]
y = df[['Duration']]
linregML(x,y, 'Humidity (%)', 'Duration (mins)', 'Effect of Humidity on Accident Duration')

Regression Score: 0.00006394


x = df[['Humidity(%)']]
y = df[['Duration']]
RANSAClinregML(x,y, 'Humidity (%)', 'Duration (mins)', 
               'Effect of Humidity on Accident Duration')

Regression Score: 0.22671241


x = df[['Wind_Chill(F)']]
y = df[['Duration']]
linregML(x,y, 'Wind Chill (F)', 'Duration (mins)', 
         'Effect of Wind Chill on Accident Duration')

Regression Score: 0.01174414


x = df[['Wind_Chill(F)']]
y = df[['Duration']]
RANSAClinregML(x,y, 'Wind Chill (F)', 'Duration (mins)', 
               'Effect of Wind Chill on Accident Duration')

Regression Score: 0.09408923


x = df[['Visibility(mi)']]
y = df[['Duration']]
linregML(x,y, 'Visibility (mi)', 'Duration (mins)', 
         'Effect of Visibility on Accident Duration')

Regression Score: 0.00015586


x = df[['Visibility(mi)']]
y = df[['Duration']]
RANSAClinregML(x,y, 'Visibility (mi)', 'Duration (mins)', 
               'The Effect of Visibility on Accident Duration')

Regression Score: 0.00294417


x = df[['Precipitation(in)']]
y = df[['Duration']]
linregML(x,y, 'Precipitation (in)', 'Duration (mins)', 
         'The Effect of Visibility on Accident Duration')

Regression Score: -0.00008211


x = df[['Precipitation(in)']]
y = df[['Duration']]
RANSAClinregML(x,y, 'Precipitation (in)', 'Duration (mins)')

Regression Score: 0.00077705


x = df[['Temperature(F)']]
y = df[['Duration']]
linregML(x,y, 'Temperature (F)', 'Duration (mins)', 
         'The Effect of Temperature on Accident Duration')

Regression Score: 0.01054072


x = df[['Temperature(F)']]
y = df[['Duration']]
RANSAClinregML(x,y, 'Temperature (F)', 'Duration (mins)', 
               'The Effect of Temperature on Accident Duration')

Regression Score: 0.05664439


numerics = ['int16', 'int32', 'int64', 'float16', 'float32', 'float64']
newdf = df.select_dtypes(include=numerics)
newdf = newdf.drop(['Start_Lat', 'End_Hour', 'Start_Hour', 'Start_Lng', 'Duration'], axis=1)
newdf


# get the features and the life expectancy data
X = newdf.values[:, 1:-1]
y = newdf.values[:, -1].astype(int)

# using random forest classification to get feature importances
clf = RandomForestClassifier(n_estimators=100, criterion="gini")
clf = clf.fit(X, y)

# set plot size and labels
plt.figure(figsize=(15, 7))
plt.title('Random forests feature importance')
plt.xlabel('Feature importance')

# create horizontal bar graph with features on the y-axis and the 
# importance of each feature as the width
plt.barh(y=newdf.columns[1:-1], width=clf.feature_importances_)
plt.show()

	ID	Severity	Start_Time	End_Time	Start_Lat	Start_Lng	Description	State	Temperature(F)	Wind_Chill(F)	...	Roundabout	Station	Stop	Traffic_Calming	Traffic_Signal	Turning_Loop	Sunrise_Sunset	Start_Hour	End_Hour	Duration
1516659	A-1516660	2	2020-12-31 09:08:00	2021-01-01 11:03:36	29.484125	-81.976435	Incident on NE 227TH PL near NE 110TH AVE Expe...	FL	71.0	71.0	...	False	False	False	False	False	False	Day	9	11	1555.600000
1638735	A-1638736	2	2020-12-31 18:04:15	2021-01-02 00:46:46	47.073525	-113.233792	Incident on UPSATA LAKE RD near WOODWORTH RD D...	MT	33.0	33.0	...	False	False	False	False	False	False	Night	18	0	1842.516667
1668370	A-1668371	2	2020-12-31 23:03:00	2021-01-02 01:01:27	32.497227	-93.735321	Incident on LISTER ST near COTY ST Drive with ...	LA	53.0	53.0	...	False	False	False	False	False	False	Night	23	1	1558.450000
1690279	A-1690280	2	2020-12-31 07:44:00	2021-01-01 10:02:57	32.601585	-93.882205	Incident on NORTHWOOD LN near WASSON RD Drive ...	LA	44.0	38.0	...	False	False	False	False	False	False	Day	7	10	1578.950000
1707259	A-1707260	2	2020-12-31 02:48:00	2021-01-01 10:02:57	32.549770	-93.773865	Incident on HAVENS near GRIMMETT DR Drive with...	LA	47.0	43.0	...	False	False	False	False	False	False	Night	2	10	1874.950000
...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...
2362639	A-2362640	2	2020-01-06 19:11:00	2020-01-06 20:26:36	35.364949	-118.967601	At CA-58/Bakersfield Tehachapi Hwy - Accident.	CA	43.0	43.0	...	False	False	False	False	False	False	Night	19	20	75.600000
2362640	A-2362641	2	2020-01-06 19:22:00	2020-01-06 19:53:38	33.775337	-117.992681	At Lampson Ave - Accident.	CA	67.0	67.0	...	False	False	False	False	False	False	Night	19	19	31.633333
2362643	A-2362644	2	2020-01-06 19:24:00	2020-01-06 20:00:38	34.139057	-118.362045	At Cahuenga Blvd - Accident.	CA	60.0	60.0	...	False	True	False	False	True	False	Night	19	20	36.633333
2362655	A-2362656	2	2020-01-06 20:49:00	2020-01-06 21:22:38	33.929463	-118.254268	At Central Ave - Accident.	CA	62.0	62.0	...	False	False	False	False	True	False	Night	20	21	33.633333
2362659	A-2362660	2	2020-01-06 21:19:00	2020-01-06 22:19:37	34.075263	-118.281157	At N Virgil Ave - Accident.	CA	64.0	64.0	...	False	False	True	True	False	False	Night	21	22	60.616667

	ID	Severity	Start_Time	End_Time	Start_Lat	Start_Lng	End_Lat	End_Lng	Distance(mi)	Description	...	Roundabout	Station	Stop	Traffic_Calming	Traffic_Signal	Turning_Loop	Sunrise_Sunset	Civil_Twilight	Nautical_Twilight	Astronomical_Twilight
42	A-43	4	2016-02-09 18:20:58	2016-02-10 00:20:58	40.451120	-85.150480	40.354290	-85.149930	6.690	Closed between IN-26 and IN-67 - Road closed d...	...	False	False	False	False	False	False	Night	Day	Day	Day
43	A-44	4	2016-02-09 18:20:58	2016-02-10 00:20:58	40.354290	-85.149930	40.451120	-85.150480	6.690	Closed between IN-67 and IN-26 - Road closed d...	...	False	False	False	False	False	False	Night	Day	Day	Day
47	A-48	4	2016-02-10 06:18:49	2016-02-10 12:18:49	40.728130	-84.789650	40.745590	-84.789620	1.206	Closed between Willshire and US-33/Rockford Rd...	...	False	False	False	False	False	False	Night	Night	Night	Day
50	A-51	2	2016-02-10 08:35:27	2016-02-10 14:35:27	41.831930	-80.101430	41.841490	-80.110990	0.824	Between Irish Rd and Blystone Rd - Accident.	...	False	False	False	False	False	False	Day	Day	Day	Day
66	A-67	2	2016-02-10 12:54:39	2016-02-10 18:54:39	41.483390	-81.662970	41.476920	-81.660750	0.462	At I-490/Exit 161 - Accident.	...	False	False	False	False	False	False	Day	Day	Day	Day
...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...
2845298	A-2845299	3	2019-08-23 16:19:21	2019-08-23 16:48:04	37.609816	-122.493720	37.608821	-122.495261	0.109	At Fassler Ave/Rockaway Beach Ave - Accident. ...	...	False	False	False	False	False	False	Day	Day	Day	Day
2845302	A-2845303	4	2019-08-23 17:25:12	2019-08-23 17:54:00	38.995930	-121.672020	39.003170	-121.662679	0.708	Closed between Yuba City and CA-99/Sawtelle Av...	...	False	False	False	False	False	False	Day	Day	Day	Day
2845303	A-2845304	4	2019-08-23 17:25:12	2019-08-23 17:54:00	39.003170	-121.662679	38.995930	-121.672020	0.708	Closed between CA-99/Sawtelle Ave and Yuba Cit...	...	False	False	False	False	False	False	Day	Day	Day	Day
2845304	A-2845305	3	2019-08-23 01:16:01	2019-08-23 01:44:53	34.121320	-117.254540	34.121320	-117.254540	0.000	At San Bernardino - Accident.	...	False	False	False	False	False	False	Night	Night	Night	Night
2845307	A-2845308	3	2019-08-23 10:32:30	2019-08-23 11:01:40	34.485189	-120.229276	34.482530	-120.230510	0.197	At Gaviota Roadside Rest Area - Accident. Lane...	...	False	False	False	False	False	False	Day	Day	Day	Day

	ID	Severity	Start_Time	End_Time	Start_Lat	Start_Lng	End_Lat	End_Lng	Distance(mi)	Description	...	Roundabout	Station	Stop	Traffic_Calming	Traffic_Signal	Turning_Loop	Sunrise_Sunset	Civil_Twilight	Nautical_Twilight	Astronomical_Twilight
246924	A-246925	2	2020-10-13 19:19:34.000000000	2021-07-19 08:00:00.000000000	44.725203	-122.185579	44.734005	-122.166744	1.107	Hazardous tree or vegetation is causing no to ...	...	False	False	False	False	False	False	Night	Night	Day	Day
250560	A-250561	2	2020-10-13 19:19:34	2021-10-01 08:00:00	44.704907	-122.231897	44.697982	-122.217820	0.841	A crash has occurred causing no to minimum del...	...	False	False	False	False	False	False	Night	Night	Day	Day
297791	A-297792	2	2020-10-13 19:19:34	2021-10-01 08:00:00	44.754713	-122.398451	44.756005	-122.391252	0.364	Rock fall has occurred causing no to minimum d...	...	False	False	False	False	False	False	Night	Night	Day	Day
312250	A-312251	2	2020-10-13 19:19:34	2021-10-01 08:00:00	44.701126	-122.074713	44.699826	-122.070964	0.205	A crash has occurred use caution.	...	False	False	False	False	False	False	Night	Night	Day	Day
320955	A-320956	2	2020-10-13 19:19:34.000000000	2021-07-19 08:00:00.000000000	44.725203	-122.185579	44.728013	-122.180288	0.324	Incident on OR-22 EB near MP 48 Drive with cau...	...	False	False	False	False	False	False	Night	Night	Day	Day
...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...
2362639	A-2362640	2	2020-01-06 19:11:00	2020-01-06 20:26:36	35.364949	-118.967601	35.364949	-118.967601	0.000	At CA-58/Bakersfield Tehachapi Hwy - Accident.	...	False	False	False	False	False	False	Night	Night	Night	Night
2362640	A-2362641	2	2020-01-06 19:22:00	2020-01-06 19:53:38	33.775337	-117.992681	33.775337	-117.992681	0.000	At Lampson Ave - Accident.	...	False	False	False	False	False	False	Night	Night	Night	Night
2362643	A-2362644	2	2020-01-06 19:24:00	2020-01-06 20:00:38	34.139057	-118.362045	34.139057	-118.362045	0.000	At Cahuenga Blvd - Accident.	...	False	True	False	False	True	False	Night	Night	Night	Night
2362655	A-2362656	2	2020-01-06 20:49:00	2020-01-06 21:22:38	33.929463	-118.254268	33.929463	-118.254268	0.000	At Central Ave - Accident.	...	False	False	False	False	True	False	Night	Night	Night	Night
2362659	A-2362660	2	2020-01-06 21:19:00	2020-01-06 22:19:37	34.075263	-118.281157	34.075263	-118.281157	0.000	At N Virgil Ave - Accident.	...	False	False	True	True	False	False	Night	Night	Night	Night

	ID	Severity	Start_Time	End_Time	Start_Lat	Start_Lng	Description	State	Temperature(F)	Wind_Chill(F)	...	Junction	No_Exit	Railway	Roundabout	Station	Stop	Traffic_Calming	Traffic_Signal	Turning_Loop	Sunrise_Sunset
246924	A-246925	2	2020-10-13 19:19:34.000000000	2021-07-19 08:00:00.000000000	44.725203	-122.185579	Hazardous tree or vegetation is causing no to ...	OR	60.0	60.0	...	False	False	False	False	False	False	False	False	False	Night
250560	A-250561	2	2020-10-13 19:19:34	2021-10-01 08:00:00	44.704907	-122.231897	A crash has occurred causing no to minimum del...	OR	60.0	60.0	...	False	False	False	False	False	False	False	False	False	Night
297791	A-297792	2	2020-10-13 19:19:34	2021-10-01 08:00:00	44.754713	-122.398451	Rock fall has occurred causing no to minimum d...	OR	60.0	60.0	...	False	False	False	False	False	False	False	False	False	Night
312250	A-312251	2	2020-10-13 19:19:34	2021-10-01 08:00:00	44.701126	-122.074713	A crash has occurred use caution.	OR	60.0	60.0	...	False	False	False	False	False	False	False	False	False	Night
320955	A-320956	2	2020-10-13 19:19:34.000000000	2021-07-19 08:00:00.000000000	44.725203	-122.185579	Incident on OR-22 EB near MP 48 Drive with cau...	OR	60.0	60.0	...	False	False	False	False	False	False	False	False	False	Night
...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...
2362639	A-2362640	2	2020-01-06 19:11:00	2020-01-06 20:26:36	35.364949	-118.967601	At CA-58/Bakersfield Tehachapi Hwy - Accident.	CA	43.0	43.0	...	False	False	False	False	False	False	False	False	False	Night
2362640	A-2362641	2	2020-01-06 19:22:00	2020-01-06 19:53:38	33.775337	-117.992681	At Lampson Ave - Accident.	CA	67.0	67.0	...	False	False	False	False	False	False	False	False	False	Night
2362643	A-2362644	2	2020-01-06 19:24:00	2020-01-06 20:00:38	34.139057	-118.362045	At Cahuenga Blvd - Accident.	CA	60.0	60.0	...	False	False	True	False	True	False	False	True	False	Night
2362655	A-2362656	2	2020-01-06 20:49:00	2020-01-06 21:22:38	33.929463	-118.254268	At Central Ave - Accident.	CA	62.0	62.0	...	False	False	False	False	False	False	False	True	False	Night
2362659	A-2362660	2	2020-01-06 21:19:00	2020-01-06 22:19:37	34.075263	-118.281157	At N Virgil Ave - Accident.	CA	64.0	64.0	...	False	False	False	False	False	True	True	False	False	Night

	Severity	Temperature(F)	Wind_Chill(F)	Humidity(%)	Pressure(in)	Visibility(mi)	Wind_Speed(mph)	Precipitation(in)
1516659	2	71.0	71.0	81.0	30.03	10.0	10.0	0.00
1638735	2	33.0	33.0	96.0	26.76	8.0	3.0	0.00
1668370	2	53.0	53.0	93.0	29.44	10.0	10.0	0.00
1690279	2	44.0	38.0	89.0	29.74	9.0	12.0	0.00
1707259	2	47.0	43.0	90.0	29.78	4.0	9.0	0.26
...	...	...	...	...	...	...	...	...
2362639	2	43.0	43.0	86.0	29.87	5.0	3.0	0.00
2362640	2	67.0	67.0	15.0	30.16	10.0	5.0	0.00
2362643	2	60.0	60.0	24.0	29.40	10.0	6.0	0.00
2362655	2	62.0	62.0	26.0	30.16	10.0	3.0	0.00
2362659	2	64.0	64.0	24.0	30.03	10.0	5.0	0.00

Analyzing US Vehicular Accidents and their Causes

Abhinav Garlipati, Ahmed Hamad, Edward Zhang, and Ved Chugh

Introduction

Data Collection

Data Processing

And the Purge Begins!!!

Analysis and Visualization

Machine Learning Models

Interpretation