Module 1, Practical 10¶

import pandas as pd
import matplotlib.pyplot as plt
import numpy as np

def get_longest_call(calls):
    longest_call = calls.loc[calls['duration'].idxmax()]
    id_caller, id_callee = longest_call.caller,longest_call.callee
    return id_caller,id_callee


def get_friends_fb(fb):
    fb_friends = []
    for _,i in fb.iterrows():
        user_a = i[0]
        user_b = i[1]
        fb_friends.append((user_a,user_b))

    return fb_friends



def get_gender(gender,id_caller,id_callee):
    gender_dict = dict()
    for _,i in gender.iterrows():
        if i[0] == 0:
            gender_dict[i[0]] = "Male"
        else:
            gender_dict[i[0]] = "Female"

    print("gender of ",id_callee," = ",gender_dict[id_callee])
    print("gender of ",id_caller," = ",gender_dict[id_caller])



def get_sms_count_dict(sms):
    sms_count = dict()
    for _,i in sms.iterrows():
        _,sender,reciver = i

        if sender > reciver:
            tmp = sender
            sender = reciver
            reciver = tmp

        if (sender,reciver) in sms_count:
            sms_count[(sender,reciver)] += 1
        else:
            sms_count[(sender,reciver)] = 1

    return sms_count


def get_tot_messages(sms_count,fb_friends):
    tot_messages = 0
    for k,v in sms_count.items():
        id_a, id_b = k

        if not (id_a,id_b) in fb_friends:
            if not (id_b,id_a) in fb_friends:
                tot_messages += v
    return tot_messages


# load data
fb = pd.read_csv("DTU/fb_friends.csv")
calls = pd.read_csv("DTU/calls.csv")
sms = pd.read_csv("DTU/sms.csv")
gender = pd.read_csv("DTU/genders.csv")


# 1)
id_caller, id_callee = get_longest_call(calls)
print("1)","the longest call is:",id_caller, "and", id_callee)

# 2)
fb_friends = get_friends_fb(fb)
if (id_callee,id_caller) in fb_friends or (id_caller,id_callee) in fb_friends:
    print("2)",id_callee," and ",id_caller," are friend of FB")



# 3)
print("3)")
get_gender(gender,id_caller,id_callee)

# 4)
sms_count = get_sms_count_dict(sms)
print("4)","they exchanged ",sms_count[(id_caller,id_callee)]," sms")


nb_messages = get_tot_messages(sms_count,fb_friends)
print("5)", nb_messages)

# make the dataset smaller
calls_small = calls[0:500]
sms_small = sms[0:5000]

def get_discretized_times(sms_small,calls_small):
    times_sms = []
    times_call = []

    for t,a,b in sms_small.to_numpy():
        times_sms.append(int(t/(60*60)))

    for t,a,b,d in calls_small.to_numpy():
        times_call.append(int(t/(60*60)))

    return times_sms, times_call

def my_plot(times_sms,times_call):
    t_s,count_s = np.unique(times_sms,return_counts=True)
    t_c,count_c = np.unique(times_call,return_counts=True)

    x_ticks = []
    x_labels = []
    for i in range(int((max(t_c)/24)+1)):
        x_ticks.append(i*24)
        x_labels.append("Day "+str(i))

    plt.figure(figsize=(15,4))
    plt.plot(t_s,count_s,label="sms",color="red")
    plt.plot(t_c,count_c,label="calls",color="blue")
    plt.ylabel("Number of interactions")
    plt.xticks(x_ticks, x_labels)
    plt.title("SMS and Calls")
    plt.legend()
    plt.savefig("nb_interactions.png")
    plt.show()


times_sms, times_call = get_discretized_times(sms_small,calls_small)
my_plot(times_sms,times_call)