Zainab Munir

IMPORT IMPORTANT LIBRARIES AND DATASET: import pandas as pd import matplotlib.pyplot as plt import seaborn as sns import numpy as np import time from IPython.display import clear_output %matplotlib inline import warnings warnings.filterwarnings('ignore') terrorism = pd.read_csv('/kaggle/input/global-terrorism-data-base/globalterrorismd b.csv', encoding='latin-1') ANALYZE DATASET: terrorism.info() RangeIndex: 181691 entries, 0 to 181690 Columns: 135 entries, eventid to related dtypes: float64(55), int64(22), object(58) memory usage: 187.1+ MB terrorism.head() eventid country \- iyear imonth 1970 7 2 NaN 0 NaN 1970 0 0 NaN 0 NaN 1970 1 0 NaN 0 NaN 1970 1 0 NaN 0 NaN 1970 1 0 NaN 0 NaN country_txt dbsource \ 0 Dominican Republic PGIS 1 Mexico PGIS 2 Philippines PGIS region iday approxdate extended resolution ... addnotes scite1 scite2 scite3 2 ... NaN NaN NaN NaN 1 ... NaN NaN NaN NaN 5 ... NaN NaN NaN NaN 3 PGIS 4 PGIS 0 1 2 3 4 INT_LOG 0 0 -9 -9 -9 Greece 8 ... NaN NaN NaN NaN Japan 4 ... NaN NaN NaN NaN INT_IDEO INT_MISC INT_ANY 0 0 0 1 1 1 -9 1 1 -9 1 1 -9 1 1 related NaN NaN NaN NaN NaN [5 rows x 135 columns] RENAME IMPORTANT COLUMNS: terrorism.rename(columns={'eventid':'Eventid', 'iyear':'Year', 'imonth':'Month', 'iday':'Day', 'extended':'Extended', 'resolution':'Resolution', 'attacktype1_txt':'Attacktype', 'country_txt':'Country', 'region_txt':'Region', 'provstate': 'Provstate', 'city':'City', 'crit1':'Crit1', 'crit2':'Crit2', 'crit3':'Crit3', 'multiple':'Multiple', 'success':'Success', 'suicide':'Suicide', 'targtype1_txt':'Targtype', 'natlty1_txt':'Natlty1', 'natlty2_txt':'Natlty2', 'natlty3_txt':'Natlty3', 'gname':'Gname', 'gname2':'Gname2', 'gname3':'Gname3', 'guncertain1':'Guncertain1', 'guncertain2':'Guncertain2', 'guncertain3':'Guncertain3', 'claimed':'Claimed', 'weaptype1_txt':'Weaptype', 'weapsubtype1_txt':'Weapsubtype', 'nkill':'Nkill', 'nkillus':'Nkillus', 'nkillter':'Nkillter', 'nwound':'Nwound', 'nwoundus':'Nwoundus', 'nwoundte':'Nwoundter', 'property':'Property', 'propextent_txt':'Propextent', 'propvalue':'Propvalue', 'ishostkid':'Ishostkid', 'nhostkid':'Nhostkid', 'nhostkidus':'Nhostkidus', 'ransom':'Ransom', 'hostkidoutcome':'Hostkidoutcome', 'nreleased':'Nreleased'},inplace=True) CREATING A NEW DATAFRAME BY SEPARATING TOP 20 IMPORTANT COLUMNS FOR ANALYSIS: data = terrorism[['Eventid', 'Year','Country', 'Region', 'Provstate', 'City', 'Crit1', 'Crit2', 'Crit3', 'Success', 'Suicide', 'Attacktype', 'Targtype', 'Natlty1','Gname', 'Guncertain1', 'Claimed', 'Weaptype', 'Nkill','Nwound']] data.head() Eventid Year Country Region \ 0 - 1970 Dominican Republic Central America & Caribbean 1 - 1970 Mexico North America 2 - 1970 Philippines Southeast Asia 3 - 1970 Greece Western Europe 4 - 1970 Japan East Asia - Provstate NaN Federal Tarlac Attica Fukouka City Santo Domingo Mexico city Unknown Athens Fukouka Crit1 1 1 1 1 1 Attacktype Assassination Hostage Taking (Kidnapping) Assassination Bombing/Explosion Facility/Infrastructure Attack Crit2 1 1 1 1 1 Crit3 1 1 1 1 1 Success 1 1 1 1 1 Suicide 0 0 0 0 0 Targtype Private Citizens & Property Government (Diplomatic) Journalists & Media Government (Diplomatic) Government (Diplomatic) \ \ Natlty1 Gname Guncertain1 \ 0 Dominican Republic MANO-D 0.0 1 Belgium 23rd of September Communist League 0.0 2 United States Unknown 0.0 3 United States Unknown 0.0 4 United States Unknown 0.0 0 1 Claimed NaN NaN Weaptype Unknown Unknown Nkill 1.0 0.0 Nwound 0.0 0.0 2 3 4 NaN NaN NaN Unknown Explosives Incendiary 1.0 NaN NaN 0.0 NaN NaN data.tail(- Eventid- Year- Provstate Suicide \ 181686 Middle Shebelle- Lattakia- Maguindanao- Manipur- Maguindanao- Claimed 1.0 0.0 0.0 Region Sub-Saharan Africa Middle East & North Africa Southeast Asia South Asia Southeast Asia \ City Crit1 Crit2 Crit3 Success Ceelka Geelow 1 1 0 1 Jableh 1 1 0 1 Kubentog 1 1 1 1 Imphal 1 1 1 0 Cotabato City 1 1 1 0 Attacktype Armed Assault Bombing/Explosion Facility/Infrastructure Attack Bombing/Explosion Bombing/Explosion Natlty1 Guncertain1 \ 181686 Somalia- Russia- Philippines- India- Philippines- Country Somalia Syria Philippines India Philippines Targtype Military Military Private Citizens & Property Government (General) Unknown Gname Al-Shabaab Muslim extremists Bangsamoro Islamic Freedom Movement (BIFM) Weaptype Firearms Explosives Incendiary Unknown Unknown Nkill 1.0 2.0 0.0 Nwound 2.0 7.0 0.0 \ - 0.0 0.0 Explosives Explosives 0.0 0.0 0.0 0.0 FIND COUNTRY WITH THE HIGHEST NUMBER OF TERRORISTS ATTACKS: print("Country with the Highest No. of Terrorist Attacks:",data['Country'].value_counts()) Country with the Highest No. of Terrorist Attacks: Country Iraq 24636 Pakistan 14368 Afghanistan 12731 India 11960 Colombia 8306 ... International 1 Wallis and Futuna 1 South Vietnam 1 Andorra 1 Antigua and Barbuda 1 Name: count, Length: 205, dtype: int64 print("Country with the Highest No. of Terrorist Attacks:",data['Country'].value_counts().index[0]) Country with the Highest No. of Terrorist Attacks: Iraq FIND REGION WITH THE HIGHEST NUMBER OF TERRORIST ATTACKS: print("Region with the Highest No. of Terrorist Attacks:",data['Region'].value_counts().index[0]) Region with the Highest No. of Terrorist Attacks: Middle East & North Africa FIND MAXIMUM NUMBER OF PEOPLE KILLED BY A SINGLE TERRORISTS ATTACK: print("Maximum No. of people killed by a single terrorist Attack are:",data['Nkill'].max()) Maximum No. of people killed by a single terrorist Attack are: 1570.0 data['Damage']=data['Nkill']+data['Nwound'] data['Damage'] 0 1 2 3 1.0 0.0 1.0 NaN 4 NaN ..- Name: Damage, Length: 181691, dtype: float64 data['Country']=data['Country'].replace('South Veitnam','Veitnam',regex=True) data['Weaptype']=data['Weaptype'].replace('Vehicle.*','Vehicle',regex= True) data['Attacktype']=data['Attacktype'].replace('Hostage Taking.*','Hostage Taking',regex=True) FIND DOMINANT ATTACK TYPE: atk_filtered=data['Attacktype'].apply(lambda x:x if x in ['Bombing/Explosion','Armed Assault','Assassination', 'Hostage Taking','Facility/Infrastructure Attack'] else 'Others') attack_type=atk_filtered.value_counts().tolist() attack_type [88255, 42669, 19312, 12149, 10356, 8950] labels=['Bombing/Explosion','Armed Assault','Assassination','Hostage Taking','Facility/Infrastructure Attack','Others'] sizes=[] for i in attack_type: percent=(i*100)/len(data['Attacktype']) sizes.append(percent) fig,ax=plt.subplots(figsize=(10,10)) patches,texts,autotexts=ax.pie(sizes,labels=labels,autopct='%1.1f% %',startangle=-20, shadow=True,explode=(0.05,0,0,0,0,0), colors=sns.color_palette('Set2',8)[:5]+ [(0.58,0.64,0.65)], textprops={'fontsize':15,'weight':'light','color':'k'}) ax.axis('equal') plt.title('Terrorist attack types',fontsize=25,pad=-70, color=sns.cubehelix_palette(8,start=.5,rot=-.75)[-3]) plt.tight_layout() ax.legend(loc='lower right',framealpha=0.5,bbox_to_anchor=(1.8,0.5,0.1,1), prop={'size':14}) fig.show() CHECK DAMAGE RATE w.r.t ATTACK TYPES: cat=["Bombing/Explosion","Armed Assault","Assassination","Hostage Taking","Facility/Infrastructure Attack"] color_cat=sns.color_palette("Set2",8)[:5] color_cat_dict=dict(zip(cat,color_cat)) color_cat_dict {'Bombing/Explosion': (0.4,-,-), 'Armed Assault': -,-,-), 'Assassination': -,-,-), 'Hostage Taking': -,-,-), 'Facility/Infrastructure Attack': -, -,-)} table1=data[["Attacktype","Damage"]].groupby("Attacktype",as_index=Fal se).sum().sort_values(by="Damage",ascending=False) table1=table1.reset_index() table- index- Attacktype Bombing/Explosion Armed Assault Unknown Assassination Hostage Taking Hijacking Unarmed Assault Facility/Infrastructure Attack Damage- inci=[8825,42669,7276,19312,12149,659,1015,10356] table1["Incidents"]=np.array(inci) table1["Damage rate"]=table1["Damage"]/table1["Incidents"] table1 index Attacktype rate 0 2 Bombing/Explosion- Armed Assault- Unknown- Assassination- Hostage Taking- Hijacking- Unarmed Assault- Facility/Infrastructure Attack- labels=table1["Attacktype"].tolist() labels ['Bombing/Explosion', 'Armed Assault', Damage Incidents - 8825 - 42669 39606.0 7276 37209.0 19312 27732.0 12149 20642.0 659 14791.0 1015 6293.0 10356 Damage 'Unknown', 'Assassination', 'Hostage Taking', 'Hijacking', 'Unarmed Assault', 'Facility/Infrastructure Attack'] x=np.arange(len(labels)) dmg=table1["Damage"].tolist() gray=(0.5,0.64,0.65) color_list=[color_cat_dict[ter_type] if ter_type in color_cat_dict.keys() else gray for ter_type in labels] d_rate=table1["Damage rate"].tolist() fig,ax1=plt.subplots(figsize=(15,8)) # Bar chart ax1.bar(labels,dmg,color=color_list,align='center') #Number in bar chart for i,v in enumerate(dmg): ax1.text(i-0.3,v-13000 if v==37209 else v+3000, str(round(v)),color='w' if v==37209 else 'k',fontweight='bold') #insert a second plot -line plot ax2=ax1.twinx() ax2.plot(labels,d_rate,linestyle='--',linewidth=4,marker='o',markerfac ecolor='black',markersize='10', label='Mean of Dead/Injured People',color='#C44D51') plt.title('Terrorist Attack Types and Damage',fontsize=25,pad=20, color=sns.cubehelix_palette(8,start=.5,rot=-.75)[-3]) ax1.set(xlabel='Types of Terrorist Attacks',ylabel='Number of Dead/Injured People') ax1.set_xticklabels(labels,rotation=45) plt.yticks(fontsize=10) ax2.legend(loc='upper center') fig.show() table2=data[["Weaptype","Damage"]].groupby("Weaptype",as_index=False). sum().sort_values(by="Damage",ascending=False) table- Weaptype Explosives Firearms Unknown Vehicle Melee Chemical Incendiary Biological Sabotage Equipment Other Radiological Fake Weapons Damage- df_count=data["Weaptype"].value_counts() df_count Weaptype Explosives 92426 Firearms 58524 Unknown 15157 Incendiary 11135 Melee 3655 Chemical 321 Sabotage Equipment 141 Vehicle 136 Other 114 Biological 35 Fake Weapons 33 Radiological 14 Name: count, dtype: int64 df_count=df_count.reindex(table2["Weaptype"].values) df_count Weaptype Explosives 92426 Firearms 58524 Unknown 15157 Vehicle 136 Melee 3655 Chemical 321 Incendiary 11135 Biological 35 Sabotage Equipment 141 Other 114 Radiological 14 Fake Weapons 33 Name: count, dtype: int64 table2["Weapcount"]=df_count.values table- Weaptype Explosives Firearms Unknown Vehicle Melee Chemical Incendiary Biological Sabotage Equipment Other Radiological Fake Weapons Damage- Weapcount- table2["Weap_Damage_rate"]=table2["Damage"]/table2["Weapcount"] table- Weaptype Explosives Firearms Unknown Vehicle Melee Chemical Incendiary Biological Sabotage Equipment Other Radiological Fake Weapons Damage- Weapcount- Weap_Damage_rate- CHECK DAMAGE RATE w.r.t WEAPON TYPES: weaplabels=table2["Weaptype"].tolist() weaplabels ['Explosives', 'Firearms', 'Unknown', 'Vehicle', 'Melee', 'Chemical', 'Incendiary', 'Biological', 'Sabotage Equipment', 'Other', 'Radiological', 'Fake Weapons'] y=np.arange(len(weaplabels)) y array([ 0, 1, 2, 3, 4, 5, 6, weapdmg=table2["Damage"].tolist() weapdmg [-,-, 50167.0, 20297.0, 14489.0, 14377.0, 8898.0, 7, 8, 9, 10, 11]) 814.0, 357.0, 225.0, 6.0, 1.0] wgray=(0.5,0.64,0.65) w_cat=["Explosives","Firearms","Unknown,Incendiary"] wcolorcat=sns.color_palette("Set2",8)[:4:1] wcolorcat [(0.4,-,-), -,-,-), -,-,-), -,-,-)] wcolor_cat_dict=dict(zip(w_cat,wcolorcat)) wcolor_cat_dict {'Explosives': (0.4,-,-), 'Firearms': -,-,-), 'Unknown,Incendiary': -,-,-)} wcolor_list=[wcolor_cat_dict[ter_type]if ter_type in wcolor_cat_dict.keys() else wgray for ter_type in weaplabels] wcolor_list [(0.4,-,-), -,-,-), (0.5, 0.64, 0.65), (0.5, 0.64, 0.65), (0.5, 0.64, 0.65), (0.5, 0.64, 0.65), (0.5, 0.64, 0.65), (0.5, 0.64, 0.65), (0.5, 0.64, 0.65), (0.5, 0.64, 0.65), (0.5, 0.64, 0.65), (0.5, 0.64, 0.65)] w_rate=table2["Weap_Damage_rate"].tolist() w_rate [-,-,-,-,-,-,-,-,-,-,-,-] PLOTTING: fig,ax1=plt.subplots(figsize=(15,8)) # Bar chart ax1.bar(weaplabels,weapdmg,color=wcolor_list,align='center') #Number in bar chart for i,v in enumerate(weapdmg): ax1.text(i-0.3,v-13000 if v==37209 else v+3000, str(round(v)),color='w' if v==37209 else 'k',fontweight='bold') #insert a second plot -line plot ax2=ax1.twinx() ax2.plot(weaplabels,w_rate,linestyle='--',linewidth=4,marker='o',marke rfacecolor='black',markersize='10', label='Mean of Dead/Injured People',color='#C44D51') plt.title('Weapon Types and Damage',fontsize=25,pad=20, color=sns.cubehelix_palette(8,start=.5,rot=-.75)[-3]) ax1.set(xlabel='Type of Weapons',ylabel='Number of Dead/Injured People') ax1.set_xticklabels(weaplabels,rotation=45) plt.yticks(fontsize=10) ax2.legend(loc='upper center') fig.show() TOP 10 TARGETED COUNTRIES FROM 2000 TO 2011: country_damage= data[data["Year"]>1999] [["Country","Damage"]].groupby("Country",as_index=False).sum() country_damage 0 1 2 3 4 .- Country Afghanistan Albania Algeria Angola Argentina ... Western Sahara Yemen Yugoslavia Zambia Zimbabwe Damage- ..- [167 rows x 2 columns] data_paint=country_damage.sort_values(by="Damage",ascending=False).hea d(10) data_paint - Country Iraq Afghanistan Pakistan India United States Nigeria Syria Yemen Somalia Philippines Damage- PLOTTING: fig,ax=plt.subplots(figsize=(12,6)) ax.barh(data_paint['Country'][::-1],data_paint['Damage'][::1],color='red') plt.xticks(rotation=-45) ax.set_ylabel('Countries',size=16) ax.set_xlabel('Number of Dead/Injured People',size=16) plt.title('Top 10 Attacked Countries From 2000 To 2011',fontsize=20,pad=10,color=sns.cubehelix_palette(8,start=.5,rot=-. 75)[-3]) plt.show() TOP 10 TARGETED NATIONALITIES FROM 2000 TO 2011: nalty_dmg=data[data["Year"]>1999] [["Natlty1","Damage"]].groupby("Natlty1",as_index=False).sum() nalty_dmg 0 1 2 3 4 .- Natlty1 Afghanistan Albania Algeria Angola Argentina ... West Bank and Gaza Strip Yemen Yugoslavia Zambia Zimbabwe Damage- ..- [186 rows x 2 columns] data["Natlty1"] 0 1 2 3 4 Dominican Republic Belgium United States United States United States ... 181686 Somalia 181687 Russia 181688 Philippines 181689 India 181690 Philippines Name: Natlty1, Length: 181691, dtype: object data_paint_natly=nalty_dmg.sort_values(by='Damage',ascending=False).he ad(10) data_paint_natly=nalty_dmg.sort_values(by="Damage",ascending=False).he ad(10) data_paint_natly- Natlty1 Iraq Afghanistan Pakistan India United States Nigeria Syria Damage- - Yemen Somalia Philippines - PLOTTING: fig,ax=plt.subplots(figsize=(12,6)) ax.barh(data_paint_natly['Natlty1'][::-1],data_paint['Damage'][::1],color='blue') plt.xticks(rotation=-45) ax.set_ylabel('Nationality',size=16) ax.set_xlabel('Number of Dead/Injured People',size=16) plt.title('Top 10 Targeted Nationalities From 2000 To 2011',fontsize=20,pad=10,color=sns.cubehelix_palette(8,start=.5,rot=-. 75)[-3]) plt.show() THE SAFEST COUNTRIES FROM 2000 TO 2017 WHICH HAVE "0" DEAD/INJURED PEOPLE CAUSED BY TERRORIST ATTACKS: country_damage[country_damage["Damage"]==0]- Country Bahamas Belize Cyprus Iceland New Zealand Damage- - Portugal Slovenia Swaziland Togo Vietnam - country_damage 0 1 2 3 4 .- Country Afghanistan Albania Algeria Angola Argentina ... Western Sahara Yemen Yugoslavia Zambia Zimbabwe Damage- ..- [167 rows x 2 columns] len(country_damage[country_damage["Damage"]==0]) 10 TIME-SERIES TOP 10 ATTACKED COUNTRIES AND TOP 10 TARGETED NATIONALITIES: def country_nationality(n): ''' Function to draw the bar charts of top 10 Attacked Countries and Top 10 Targeted Nationalities''' #Filter data and choose top 10 Attacked Counties top10_bad=data[data['Year']==n] [['Country','Damage']].groupby('Country',as_index=False).sum() data_paint=top10_bad.sort_values(by='Damage',ascending=False).head(10) fig,ax=plt.subplots(1,2,figsize=(16,4)) ax[0].barh(data_paint['Country'][::-1],data_paint['Damage'][::-1], color=['red','green','#344650','#df2029','#FFFC00', '#E4405F','#3B5999','#00c300','#ff0084']) ax[0].set_ylabel('Countries',size=16) ax[0].set_xlabel('Number of Dead/Injured People',size=16) ax[0].set_title('Top 10 Attacked Countries %d' %n,fontsize=15,pad=10,color=sns.cubehelix_palette(8,start=.5,rot=-.75) [-3]) plt.subplots_adjust(wspace=0.3) #Choose column Country and Damage from data nalty_dmg=data[data['Year']>1999] [['Natlty1','Damage']].groupby('Natlty1',as_index=False).sum() data_paint_natly=nalty_dmg.sort_values(by='Damage',ascending=False).he ad(10) ax[1].barh(data_paint_natly['Natlty1'][::-1],data_paint['Damage'] [::-1],color='blue') ax[1].set_ylabel('Nationality',size=16) ax[1].set_xlabel('Number of Dead/Injured People',size=16) ax[1].set_title('Top 10 Targeted Nationalities %d' %n,fontsize=20,pad=10,color=sns.cubehelix_palette(8,start=.5,rot=-.75) [-3]) plt.show() for ax in fig.axes: plt.sca(ax) plt.show() list_year=[] for year in range(1970,1993): list_year.append(year) for year in range(1994,2012): list_year.append(year) for year in list_year: country_nationality(year) time.sleep(2) clear_output() THE TOP 10 ATTACKED COUNTRIES AND TOP 10 TARGETED NATIONALITIES WITH THE SPECIFIC YEAR INPUT: def country_nationality_year(): ''' Function to paint chart with specific input year''' # Input year. Try if in list year try: year=int(input('Input the year you want to see chart! -): ')) if year==1993: print('Sorry, No Terrorism Data in 1993') year=int(input('Input the year you want to see chart! -): ')) while year not in list_year: print('Please input the year from 1970 to 2011') year=int(input('Input the year you want to see chart! -): ')) except (ValueError,ZeroDivisionError) as ex: print('Wrong Type, Input Again: ') country_nationality_year() else: country_nationality(year) country_nationality_year()