import matplotlib.pyplot as plt
import pandas as pd
from sodapy import Socrata
import seaborn as sns
import gmaps
from ipywidgets.embed import embed_minimal_html

username = input("Enter the Username") # Prompts the user credentials for the API
password = input("Enter the password") 
MyAppToken = input('Enter the app token')
client = Socrata('data.cityofnewyork.us', MyAppToken, username=username,password=password)
results = client.get("h9gi-nx95", limit=100000) # get() pulls the dynamic data from the API.
results_df = pd.DataFrame.from_records(results) 

results_df['year'] = pd.DatetimeIndex(results_df['crash_date']).year # year, month, weekday and hour features are extracted from crash_date
results_df['month'] = pd.DatetimeIndex(results_df['crash_date']).month
results_df['weekday'] = pd.DatetimeIndex(results_df['crash_date']).weekday
results_df['hour'] = pd.DatetimeIndex(results_df['crash_time']).hour
df1=results_df[(results_df.year >= 2019)] # The crash data on and after 2019 year is used for visualization

#---------------------Bar plot------------------------------------------------

newframe = pd.DataFrame(df1.groupby(['contributing_factor_vehicle_1'])['number_of_persons_killed'].count()) # Counts the number of Accidents
newframe1 = pd.DataFrame(df1.groupby(['contributing_factor_vehicle_1'])['number_of_persons_injured'].sum()) #Counts the number of persons injured
newframe2 = pd.merge(newframea, newframe1, on='Cause') # Merges the top contributing factors' accident counts and their respective number of injuries
newframe2 = newframe2.head(5) # Picks the top contributing factors
newframe2 = newframe2.rename(columns = {"number_of_persons_injured":"No. of Persons Injured", "number_of_persons_killed":"No. of Accidents"})
fig,(ax1,ax2) =plt.subplots(2,1,figsize = (12,12))   # Fixes the size and subplot place
ax = newframe2.plot.bar(rot=0,ax=ax1, width = 0.7)   # Plots the stacked bar plot 
ax.legend(fontsize = 14)
ax.xaxis.label.set_size(14)
ax.set_title('Top Five Accident Contributing Factors', fontsize = 15)
plt.xticks(fontsize = 9, wrap = True)

#---------------------Pie plot-------------------------------------------------
newframe3 = pd.DataFrame(df1.groupby(['contributing_factor_vehicle_1'])['number_of_persons_killed'].sum()) # Sums the number of persons killed
Others = sum(newframe3['number_of_persons_killed']==1) # Adds the killed values that is equal to 1 to "Other" contributing factor
newframe3=newframe3[newframe3['number_of_persons_killed']>1] # Picks the contributing factor where the killed is greater than 1
newframe3 = newframe3.append({'number_of_persons_killed':Others,'Cause':'Others'},ignore_index=True) # Adds the "Other" to the rows
newframe3 = newframe3.head(7) # Picks only top seven contributing factor
newframe3['number_of_persons_killed'] = round((newframe3['number_of_persons_killed']/a)*100,0) # Calculates the percetage of death by each contributing factor in total kills

ax2.pie(newframe3['number_of_persons_killed'],labels=newframe3['Cause'],autopct='%1.1f%%') # Plots the Pie chart
ax2.set_title('% of Mortality by each Factor', fontsize = 15)
plt.savefig('Bar_Pie.png')

plt.savefig('Bar_Pie.png')	
plt.show()

heatmap_df1["Month"] = pd.Categorical(heatmap_df1["Month"], heatmap_df1.Month.unique()) 
plt.figure(figsize = (15, 9)) # Assigning figure size (length & breadth) for the plot
file_long = heatmap_df1.pivot("Weekday", "Month", "Crash_Count") # Assigning the column names for which the heatmap needs to be plotted
sns.heatmap(file_long, cmap = 'viridis', annot=True, fmt=".0f") # Plotting the map
plt.title("Heatmap of Crash Count in New York City (Monthly vs Weekly)", fontsize = 14); # Assigning title for the plot
plt.savefig('Heatmap1.jpg') # Saving the plot

heatmap_df2["Weekday"] = pd.Categorical(heatmap_df2["Weekday"], heatmap_df2.Weekday.unique())
plt.figure(figsize = (20, 10))
file_long = heatmap_df2.pivot("Weekday", "Hour", "Crash_Count")
sns.heatmap(file_long, cmap = 'viridis', annot=True, fmt=".0f")
plt.title("Heatmap of Crash Count in New York City (Weekly vs Hourly)", fontsize = 14);
plt.savefig('Heatmap2.jpg')

#------------------------Visualizing using Gmaps-------------------------------

locations=pd.DataFrame(results_df[['latitude','longitude']])
locations[['latitude','longitude']] = locations[['latitude','longitude']].astype(float) #Latitude and Longitude data are stored as float

gmaps.configure(api_key='Key Here') #GMAPS API key is inserted
nyc_coordinates = (40.7128, -74.0060)
fig = gmaps.figure(center=nyc_coordinates, zoom_level=10.5) #Map co-ordinates along with zoom level is set
heatmap_layer=gmaps.heatmap_layer(locations) #heatmap layer is created using latitude,longitude
heatmap_layer.max_intensity = 200
heatmap_layer.point_radius = 15
fig.add_layer(heatmap_layer)
embed_minimal_html('Heatmap_layer.html', views=[fig]) #heatmap file is exported in save directory