Google Trends

Published in

Intro to Python WOWs

19 min readOct 30, 2019

by Gregorio Cioppa and Betty Staccoli

Overview

The purpose of this project is to provide a program able to access Google Trends data in an easier and more standardizable way. In order to do this, it has been coded as a tool providing details of searches around a specific word provided by the user. These details include evolution of the searches over time, list of countries where the word has been googled the most, list of related topics and queries. This will represent the “base” MVP.

It has been possible to expand and add more flexibility to the “base” version, achieving an “improved” MVP. These improvements include additional functionalities that allow the user to set the timespan to consider, the country of interest and restrict the research to a specific category (e.g. “Art & Entertainment”).

Setup

Before starting, some specific packages has been installed, specifically:

pip install pytrends : unofficial API for Google Trends containing the functions necessary to perform the seraches
pip install pandas : API necessary to manage DataFrames
pip install matplotlib : API necessary to plot charts
Once this initial step has been performed, all the libraries necessary to run the code were imported.

Libraries to use Google Trends APIs (pytrends) include:

from pytrends.request import TrendReq
from pandas.io.json import json_normalize

Library to export dataframes as csv files includes:

import pandas as pd

Libraries for graphs and dates formatting include:

import matplotlib.pyplot as plt
import datetime
from datetime import date
import matplotlib.dates as mdates

Inputs

In this chapter all the inputs required by the user are included. It has to be noted that also the inputs for additional functionalities, such as a variable timeframe of analysis, are included here.

Step 1: Key words selection

The first step to be performed is to acquire from the user the keywords to search (up to 3).

kw_1 = input(“Please write the first keyword: “)

Please write the first keyword: Cartier

print(kw_1)

Cartier

q1 = input(“Do you want to add another keyword? [Yes/No]: “)

Do you want to add another keyword? [Yes/No]: yes

print(q1)

yes

if q1.upper() == “YES”:
    kw_2 = input(“Please write the second keyword: “)
else:
    kw_2 = “”

Please write the second keyword: Dior

if q1.upper() == “YES”:
    q2 = input(“Do you want to add another keyword? [Yes/No]: “)

Do you want to add another keyword? [Yes/No]: yes

if q2.upper() == “YES”:
    kw_3 = input(“Please write the third keyword: “)
else:
    kw_3 = “”

Please write the third keyword: Chanel

user_kw_list = [kw_1]
if q1.upper() == “YES”:
    user_kw_list.append(kw_2)
if q2.upper() == “YES”:
    user_kw_list.append(kw_3)

This is the final list of keywords that will be used to perform analyses on Google Trends.

print(user_kw_list)

[‘Cartier’, ‘Dior’, ‘Chanel’]

Step 2: Refine the research (Improved MVP)

The unofficial API for Google Trends relies on some default setting, that are used if not specified by the user. Specifically, these settings include:

Category of focus: all
Timeframe of analysis: last 5 years
Country of focus: worldwide

All the three variables have been then initialized to these values.

user_cat = 0 
user_timeframe = 'today 5-y' 
user_geo = ''

Focus on a specific category

As previously mentioned, it is possible to focus the research on a specific category of interest. In order to do this, it has been asked to the user whether she/he wants to limit the analyses first, and then to which category. All the categories are identified by a code and a coding table is provided to the user to proceed with the choice.

ans = input (f'Would you like to focus the research on a specific category? (e.g. Art & Entertainment) (Yes/No): ')

Would you like to focus the research on a specific category? (e.g. Art & Entertainment) (Yes/No): yes

if ans.upper() == "YES":
    category_list = 'Here is the list of possible categories: \n Arts & Entertainment: 3 \n Autos & Vehicles: 47 \n Beauty & Fitness: 44 \n Books & Literature: 22 \n Business & Industrial: 12 \n Computers & Electronics: 5 \n Finance: 7 \n Food & Drink: 71 \n Games: 8 \n Health: 45 \n Hobbies & Leisure: 65 \n Home & Garden: 11 \n Internet & Telecom: 13 \n Jobs & Education: 958 \n Law & Government: 19 \n News: 16 \n Online Communities: 299 \n People & Society: 14 \n Pets & Animals: 66 \n Real Estate: 29 \n Reference: 533 \n Science: 174 \n Shopping: 18 \n Sports: 20 \n Travel: 67'
    print (category_list)
    user_cat = int(input ('Please indicate the code of the category to focus on: '))

Here is the list of possible categories:
Arts & Entertainment: 3
Autos & Vehicles: 47
Beauty & Fitness: 44
Books & Literature: 22
Business & Industrial: 12
Computers & Electronics: 5
Finance: 7
Food & Drink: 71
Games: 8
Health: 45
Hobbies & Leisure: 65
Home & Garden: 11
Internet & Telecom: 13
Jobs & Education: 958
Law & Government: 19
News: 16
Online Communities: 299
People & Society: 14
Pets & Animals: 66
Real Estate: 29
Reference: 533
Science: 174
Shopping: 18
Sports: 20
Travel: 67
Please indicate the code of the category to focus on: 18

To be sure to gather a correct input, it has been introduced a check on the category code inserted by the user.

acceptable_cat = [3, 47, 44, 22, 12, 5, 7, 71, 8, 45, 65, 11, 13, 958, 19, 16, 299, 14, 66, 29, 533, 174, 18, 20, 67]
if ans.upper() == "YES":
    if user_cat in acceptable_cat:
        print ('Category set')
    else :
        user_cat = int(input ('Category code not valid, please indicate the right code: '))   
        if user_cat in acceptable_cat:
           print ('Category set')
        else :
            user_cat = int(input ('Category code not valid, please indicate the right code: '))
            if user_cat in acceptable_cat:
                print ('Category set')
            else:
                user_cat = int(input ('Category code not valid, please indicate the right code: '))

Category set

Timeframe

The second degree of flexibility that this code offers is related to the timeframe to consider. It is given to the user the possibility to customize the time range of interest, from the 5 to 1 most recent years.

ans = input ('Would you like to focus the research on a specific timeframe? (default is last 5 years) (Yes/No): ')

Would you like to focus the research on a specific timeframe? (default is last 5 years) (Yes/No): yes

if ans.upper() == "YES":
    timeframe_list = ('Here is the list of the timeframes available: \n all: all \n last 5 years: today 5-y \n last 4 years: today 4-y \n last 3 years: today 3-y \n last 2 years: today 2-y \n last 1 years: today 1-y')
    print (timeframe_list)
    user_timeframe = input ('Please indicate the timeframe of focus in the format specified in the list (e.g. "today 3-y"): ')

Here is the list of the timeframes available:
all: all
last 5 years: today 5-y
last 4 years: today 4-y
last 3 years: today 3-y
last 2 years: today 2-y
last 1 years: today 1-y
Please indicate the timeframe of focus in the format specified in the list (e.g. “today 3-y”): today 2-y

As in the previous case, in order to be sure to gather an acceptable input, it has been introduced a check.

acceptable_timeframe = ['today 5-y', 'today 4-y', 'today 3-y', 'today 2-y', 'today 1-y']
if user_timeframe in acceptable_timeframe:
    print ('Timeframe set')
else:
    user_timeframe = input ('Timeframe not valid, please indicate the right timeframe: ')    
    if user_timeframe in acceptable_timeframe:
        print ('Timeframe set')
    else:
        user_timeframe = input ('Timeframe not valid, please indicate the right timeframe: ')
        if user_timeframe in acceptable_timeframe:
            print ('Timeframe set')
        else:
            user_timeframe = input ('Timeframe not valid, please indicate the right timeframe: ')

Timeframe set

Once the input about the timeframe is received, it is coded in a new numerical variable, easier to be used in the following steps of the code (“numyears”)

if user_timeframe == "today 5-y":
    numyears = 5
elif user_timeframe == "today 4-y":
    numyears = 4
elif user_timeframe == "today 3-y":
    numyears = 3
elif user_timeframe == "today 2-y":
    numyears = 2
else:
    numyears = 1

API for Google Trends requires a specific format for the timeframe, specifically “YYYY-MM-DD YYYY-MM-DD”, where the first date is the starting date and the second one the date of today. Therefore, the input received by the user is re-formatted in this way.

Please note that this step resulted to be quite tricky as it was not clear at the beginning neither the format required for the input nor how to re-format the date in the desired form.

numdays = 7
numweeks = 52
total_time_range = numdays * numweeks * numyearsend_date= date.today()
today_date = date.today()
end_date = today_date
begin_date = end_date - datetime.timedelta(days = total_time_range-7)user_timeframe = begin_date.strftime('%Y-%m-%d')+' '+end_date.strftime('%Y-%m-%d')
print(user_timeframe)

Geographies

The third and last degree of flexibility available in the API is related to the specific country of focus. As in the previous cases, an eventual input from the user is asked for a better customization of the search.

It has to be noted that, on the contrary of the categories, it has not been possible to find a coding system for the different countries available. Yet, the official rule provided by the creator of the API is to indicate the country with the first two letters in capital (e.g. Italy, IT, United States, US).

ans = input ('Would you like to focus the research on a specific country? (default is World) (Yes/No): ')

Would you like to focus the research on a specific country? (default is World) (Yes/No): no

if ans.upper() == “YES”:
 user_geo = input (‘Please indicate country of focus (first two letters in capitals, e.g. United States is US): ‘)print(user_geo)

Google Trends outputs

Connection to Google

First, it is necessary to connect to Google. This action is required to be run just once, as the following analyses use the same session.

pytrend = TrendReq()

Secondly, it has to be created the payload, containing the settings for the analyses and capture the API tokens. These settings will be used for the analyses of interest over time, by region and the related queries.

Whilst default settings are typically used, the current Improved MVP allows to have a better customization for what regards categories, timeframe and country of focus.

pytrend.build_payload(kw_list=user_kw_list, cat=user_cat, timeframe=user_timeframe, geo=user_geo)

Interest over time

The first output is the evolution of the interest over time of the selceted keywords. This output is provided on weekly basis for the entire timeframe of analysis.

As a general note, this and all the following outputs are provided on relative basis:

Between the different keywords
Within each keywords
The maximum level of interest is equal to 100, whilst the minimum to 0

The code below provided the output and an example of the first lines.

interest_over_time_df = pytrend.interest_over_time()
print(interest_over_time_df.head())

Cartier Dior Chanel isPartial
date
2017–11–05 32 11 66 False
2017–11–12 31 11 67 False
2017–11–19 37 13 76 False
2017–11–26 35 12 73 False
2017–12–03 35 12 76 False

Rather than having the data stored in a Dataframe, the real value added is the export to a csv file.

It has to be noted that this step presented some challenges as we wanted to keep the information about the timestamp of the data, which is an index and not a column. In order to do this, the parameter “Index” had to be set to True.

interest_over_time_df.to_csv("interest_over_time.csv", index=True)

Visualization is an important aspect when dealing with many data. For this reason, a focus has been put into creating a chart.

Also this step resulted quite challenging as it was a new python function to be explored. The first issue to be faced was related to the definition of the timeline (i.e. x axis) from the start date to today. This timeline was supposed to present a 1-to-1 correspondence with the interest, yet doing calculation with dates required quite effort. Another issue was related to the plot of the interests of the different keywords as they were stored in a DataFrame. After a detailed research, it was understood that the extension .iloc allows to consider specific elements within a Dataframe and used accordingly. An additional complexity was related to the fact that multiple lines were supposed to be plotted on the same graph, all with the same scale. Finally, some formatting was made necessary to make the chart easy to be read. In this case, it was decided to adjust the scale according to the length of the timeframe.

# x axis values: dates of the last n years (n defined above by the user)
numdays = 7
numweeks = 52datelist = [begin_date]for index in range (0, numweeks*numyears-3):
    datelist.append(datelist[index] + datetime.timedelta(days = numdays)) 
x_axis = datelist# Plot into a chart 
plt.gca().xaxis.set_major_formatter(mdates.DateFormatter('%m/%Y'))
plt.gca().xaxis.set_major_locator(mdates.DayLocator(0,14*numyears))
plt.gcf().autofmt_xdate()plt.plot (x_axis, interest_over_time_df.iloc[:,0], label = kw_1)
if q1.upper() == "YES":
    plt.plot (x_axis, interest_over_time_df.iloc[:,1], label = kw_2)
if q2.upper() == "YES":
    plt.plot (x_axis, interest_over_time_df.iloc[:,2], label = kw_3)plt.xticks(rotation=45)plt.xlabel('Time')
plt.ylabel('Level of Interest')
plt.title('Interest over time')plt.legend()
plt.show()

Average searches

An relevant data to compare is the average value of the interest of all the keywords within the timeframe of analysis.

avg_kw = interest_over_time_df.mean(0)
print(f"The average interest in the last ",numyears,"years is ", avg_kw)

The average interest in the last 2 years is Cartier 35.137255
Dior 15.588235
Chanel 70.647059
dtype: float64

Interest by Region

The second output provided in the interest by region in the timeframe considered. Also in this case, the value provided is relative both within the countries and between the selected keywords. For the clarity of analysis, the regions with low interested (i.e. about 0) have been removed from the file.

The level of Region depends on whether a variable for country has been inserted by the user. In fact, if the user has not refine the search by country, the output by region will refer to the list of worldwide countries. If a country has been selected, the output by region will show the regions or states for that country.

interest_by_region_df = pytrend.interest_by_region(inc_low_vol=False)
print(interest_by_region_df.head())

Cartier Dior Chanel
geoName
Afghanistan 0 0 0
Albania 0 0 0
Algeria 16 35 49
American Samoa 0 0 0
Andorra 0 0 0

The output obtained is exported to a csv file:

interest_by_region_df.to_csv(“interest_by_region.csv”, index=True)

A good way to look at this data is, for each of the keyword, to rank the countries in descendent order, in order to understand with are the countries registering the most and the least interest on the keyword.

As additional note, in case US is selected, the code provided the interest in the different states, in case of Italy, the details of the region.

print('RANKED COUNTRIES FOR KEYWORD "',kw_1,'"' )
print("")
print(interest_by_region_df.sort_values(by=kw_1, ascending=False)[kw_1])

RANKED COUNTRIES FOR KEYWORD “ Cartier “

geoName
Canada 45
Romania 36
Kuwait 29
United Arab Emirates 29
Qatar 28
Saudi Arabia 27
Morocco 26
Switzerland 24
Belgium 22
France 21
Lebanon 21
Colombia 20
Mexico 19
Philippines 18
Singapore 17
United States 17
Argentina 16
Netherlands 16
Azerbaijan 16
Italy 16
Algeria 16
Tunisia 15
Hong Kong 15
Egypt 15
Kazakhstan 14
Israel 14
South Korea 14
United Kingdom 14
Spain 13
Russia 13
..
Martinique 0
Mauritania 0
Jamaica 0
Iraq 0
Ethiopia 0
Iceland 0
Falkland Islands (Islas Malvinas) 0
Faroe Islands 0
Fiji 0
French Guiana 0
French Polynesia 0
French Southern Territories 0
Gabon 0
Gambia 0
Georgia 0
Ghana 0
Gibraltar 0
Greenland 0
Grenada 0
Guadeloupe 0
Guam 0
Guatemala 0
Guernsey 0
Guinea 0
Guinea-Bissau 0
Guyana 0
Haiti 0
Heard & McDonald Islands 0
Honduras 0
Åland Islands 0
Name: Cartier, Length: 250, dtype: int32

if q1.upper() == "YES":
    print('RANKED COUNTRIES FOR KEYWORD "',kw_2,'"' )
    print("")
    print(interest_by_region_df.sort_values(by=kw_2, ascending=False)[kw_2])

RANKED COUNTRIES FOR KEYWORD “ Dior “

geoName
Japan 42
Kazakhstan 42
Taiwan 40
Russia 40
Czechia 40
Ukraine 40
France 37
Tunisia 37
Greece 37
Bulgaria 37
Poland 36
Israel 35
Algeria 35
Nigeria 34
Azerbaijan 34
Vietnam 34
Finland 33
Lebanon 33
Turkey 33
Saudi Arabia 33
Qatar 33
Morocco 33
Egypt 32
Norway 31
Serbia 31
Italy 31
South Korea 31
United Arab Emirates 30
Germany 30
Kuwait 30
..
Martinique 0
Mauritania 0
Jamaica 0
Iraq 0
Ethiopia 0
Iceland 0
Falkland Islands (Islas Malvinas) 0
Faroe Islands 0
Fiji 0
French Guiana 0
French Polynesia 0
French Southern Territories 0
Gabon 0
Gambia 0
Georgia 0
Ghana 0
Gibraltar 0
Greenland 0
Grenada 0
Guadeloupe 0
Guam 0
Guatemala 0
Guernsey 0
Guinea 0
Guinea-Bissau 0
Guyana 0
Haiti 0
Heard & McDonald Islands 0
Honduras 0
Åland Islands 0
Name: Dior, Length: 250, dtype: int32

if q2.upper() == “YES”:
    print(‘RANKED COUNTRIES FOR KEYWORD “‘,kw_3,’”’ )
    print(“”)
    print(interest_by_region_df.sort_values(by=kw_3, ascending=False)[kw_3])

RANKED COUNTRIES FOR KEYWORD “ Chanel “

geoName
Indonesia 74
Venezuela 73
Pakistan 72
Peru 72
Brazil 70
Ireland 67
India 67
Hungary 66
Denmark 66
United States 65
Chile 65
South Africa 65
Colombia 64
Thailand 64
Australia 64
Croatia 62
Portugal 62
Philippines 62
New Zealand 62
Mexico 61
Sweden 61
Singapore 61
Norway 61
Poland 60
Malaysia 60
Iran 60
Hong Kong 60
Austria 60
Vietnam 60
Serbia 59
..
Martinique 0
Mauritania 0
Jamaica 0
Iraq 0
Ethiopia 0
Iceland 0
Falkland Islands (Islas Malvinas) 0
Faroe Islands 0
Fiji 0
French Guiana 0
French Polynesia 0
French Southern Territories 0
Gabon 0
Gambia 0
Georgia 0
Ghana 0
Gibraltar 0
Greenland 0
Grenada 0
Guadeloupe 0
Guam 0
Guatemala 0
Guernsey 0
Guinea 0
Guinea-Bissau 0
Guyana 0
Haiti 0
Heard & McDonald Islands 0
Honduras 0
Åland Islands 0
Name: Chanel, Length: 250, dtype: int32

A visual output of top countries is provided below. In this case, an histogram results to be the most appropriate chart.

However, this visualization is most suited for cases where the user has selected a specific country so that the graph will show the ranked regions/states of that country.

print(“Interest by region”)
print(“”)
interest_by_region_df.sort_values(kw_1, ascending=False)[[kw_1]].plot.bar(stacked=True)
if q1.upper() == “YES”:
    interest_by_region_df.sort_values(kw_2, ascending=False)[[kw_2]].plot.bar(stacked=True)
if q2.upper() == “YES”:
    interest_by_region_df.sort_values(kw_3, ascending=False)[[kw_3]].plot.bar(stacked=True)

Related Queries

The third output provided by the code is about the related queries. In other words, the other words commonly related to the keywords of analysis.

related_queries_dict = pytrend.related_queries()
print(related_queries_dict)

{‘Cartier’: {‘top’: query value
0 cartier watch 100
1 watch 99
2 cartier glasses 64
3 cartier watches 39
4 cartier tank 34
5 cartier ring 26
6 cartier diamond 21
7 santos cartier 21
8 tiffany 20
9 cartier rings 20
10 rolex 20
11 cartier sunglasses 19
12 cartier bracelet 18
13 cartier men 17
14 cartier tank watch 17
15 cartier mens watch 11
16 cartier jewelry 10
17 cartier frames 9
18 rolex watch 9
19 gold cartier watch 9
20 cartier engagement rings 9
21 cartier necklace 9
22 cartier earrings 8
23 cartier band 8
24 cartier replica 8, ‘rising’: query value
0 cartier bardi 36900
1 cartier dopebwoy 16800
2 cartier bazzi 15950
3 cartier bardi lyrics 15950
4 cartier bartier 12300
5 cartier bazzi lyrics 11100
6 cartier santos 2018 10050
7 clash de cartier 8700
8 cartier dopebwoy lyrics 6300
9 cartier french tank watch two tone 4050
10 cartier glasses they sit on my face 3500
11 cartier gucci scarf lyrics 2850
12 cartier french tank watch 900
13 cardi b cartier 650
14 cartier clash 600
15 bust down cartier 450
16 cartier gucci scarf 400
17 cartier halo tiara 350
18 cartier vision 300
19 cartier ring size chart 200
20 cartier rimless glasses 200
21 diamond cartier glasses 200
22 cartier tiara 180
23 cartier store near me 180
24 tiffanys jewelry 180}, ‘Dior’: {‘top’: query value
0 christian dior 100
1 dior sunglasses 77
2 dior sneakers 61
3 dior shoes 51
4 dior bag 49
5 chanel 39
6 gucci 37
7 dior homme 27
8 dior men 25
9 vintage dior 24
10 louis vuitton 23
11 dior glasses 20
12 dior handbags 20
13 dior bags 18
14 dior bracelet 16
15 dior shirt 14
16 dior earrings 14
17 dior nordstrom 13
18 lady dior 13
19 prada 13
20 dior tote 13
21 fendi 13
22 dior belt 12
23 dior boots 12
24 chanel bag 11, ‘rising’: query value
0 dior book tote 53950
1 dior woven bracelet 40950
2 dior x kaws 34250
3 dior b23 31650
4 gianna dior 30250
5 dior b22 sneakers 28700
6 joy dior 28350
7 iann dior 24650
8 christian dior mens sneakers 22800
9 dior chucks 20450
10 dior b22 sneaker 19800
11 amiri jeans 18700
12 dior friendship bracelets 14500
13 christian dior friendship bracelet 14450
14 dior kaws shirt 14350
15 dior oblique sneakers 13800
16 christian dior headband 13650
17 dior chuck taylor 11600
18 dior messenger bag 11200
19 dior clogs 9400
20 christian dior converse 9200
21 dior padlock necklace 8850
22 christian dior visor 8100
23 dior bifocals 7850
24 christian dior woven bracelet 7350}, ‘Chanel’: {‘top’: query value
0 chanel bag 100
1 chanel bags 57
2 gucci 38
3 chanel shoes 30
4 chanel sneakers 23
5 louis vuitton 22
6 chanel purse 22
7 chanel handbags 20
8 chanel earrings 18
9 chanel wallet 17
10 chanel flap bag 14
11 chanel boy bag 13
12 chanel classic 13
13 chanel sunglasses 13
14 chanel tote 12
15 chanel necklace 11
16 vintage chanel 11
17 chanel belt 11
18 chanel designer 10
19 coco chanel 10
20 chanel mini 10
21 nordstrom chanel 10
22 chanel santini 9
23 chanel men 9
24 gucci bag 9, ‘rising’: query value
0 chanel bags 2019 23100
1 chanel sneakers 2018 13200
2 chanel sandals 2019 10150
3 chanel tote 2018 7100
4 chanel sunglasses 2018 6750
5 chanel milk carton bag 5200
6 chanel bags 2018 3100
7 chanel human race 2800
8 chanel handbags 2018 1650
9 biggest chanel bag 1350
10 chanel pharrell shoes 850
11 chanel pharrell 750
12 gucci fanny pack 250
13 chanel sneakers women 200
14 chanel belt bag 190
15 chanel pvc bag 180
16 chanel fanny pack 160
17 chanel belt womens 150
18 men chanel sneakers 150
19 chanel bucket bag 140
20 chanel boyfriend bag 140
21 audrey hepburn 130
22 chanel belt 120
23 chanel suspenders 100
24 men chanel shoes 100}}

The data exported to csv, in this case, are very difficult to read, as shown in the output below.

(pd.DataFrame.from_dict(data=related_queries_dict, orient='index')
   .to_csv('related_queries_dict.csv', header=True))

For this reason, it has been decided to divide the dictionary into different outputs, one per keywords, that are much easier to read and intepret. For each keywords, details of the Top and rising queries are provided separately.

print(‘keyword = “ ‘, kw_1,’ “ Top queries’)
print(“”)
print(related_queries_dict[kw_1][“top”])

keyword = “ Cartier “ Top queries

query value
0 cartier watch 100
1 watch 99
2 cartier glasses 64
3 cartier watches 39
4 cartier tank 34
5 cartier ring 26
6 cartier diamond 21
7 santos cartier 21
8 tiffany 20
9 cartier rings 20
10 rolex 20
11 cartier sunglasses 19
12 cartier bracelet 18
13 cartier men 17
14 cartier tank watch 17
15 cartier mens watch 11
16 cartier jewelry 10
17 cartier frames 9
18 rolex watch 9
19 gold cartier watch 9
20 cartier engagement rings 9
21 cartier necklace 9
22 cartier earrings 8
23 cartier band 8
24 cartier replica 8

print(‘keyword = “ ‘, kw_1,’ “ Rising queries’)
print(“”)
print(related_queries_dict[kw_1][“rising”])

keyword = “ Cartier “ Rising queries

query value
0 cartier bardi 36900
1 cartier dopebwoy 16800
2 cartier bazzi 15950
3 cartier bardi lyrics 15950
4 cartier bartier 12300
5 cartier bazzi lyrics 11100
6 cartier santos 2018 10050
7 clash de cartier 8700
8 cartier dopebwoy lyrics 6300
9 cartier french tank watch two tone 4050
10 cartier glasses they sit on my face 3500
11 cartier gucci scarf lyrics 2850
12 cartier french tank watch 900
13 cardi b cartier 650
14 cartier clash 600
15 bust down cartier 450
16 cartier gucci scarf 400
17 cartier halo tiara 350
18 cartier vision 300
19 cartier ring size chart 200
20 cartier rimless glasses 200
21 diamond cartier glasses 200
22 cartier tiara 180
23 cartier store near me 180
24 tiffanys jewelry 180

if q1.upper() == "YES":
    print('keyword = " ', kw_2,' " Top queries')
    print("")
    print(related_queries_dict[kw_2]["top"])

keyword = “ Dior “ Top queries

query value
0 christian dior 100
1 dior sunglasses 77
2 dior sneakers 61
3 dior shoes 51
4 dior bag 49
5 chanel 39
6 gucci 37
7 dior homme 27
8 dior men 25
9 vintage dior 24
10 louis vuitton 23
11 dior glasses 20
12 dior handbags 20
13 dior bags 18
14 dior bracelet 16
15 dior shirt 14
16 dior earrings 14
17 dior nordstrom 13
18 lady dior 13
19 prada 13
20 dior tote 13
21 fendi 13
22 dior belt 12
23 dior boots 12
24 chanel bag 11

if q1.upper() == "YES":
    print('keyword = " ', kw_2,' " Rising queries')
    print("")
    print(related_queries_dict[kw_2]["rising"])

keyword = “ Dior “ Rising queries

query value
0 dior book tote 53950
1 dior woven bracelet 40950
2 dior x kaws 34250
3 dior b23 31650
4 gianna dior 30250
5 dior b22 sneakers 28700
6 joy dior 28350
7 iann dior 24650
8 christian dior mens sneakers 22800
9 dior chucks 20450
10 dior b22 sneaker 19800
11 amiri jeans 18700
12 dior friendship bracelets 14500
13 christian dior friendship bracelet 14450
14 dior kaws shirt 14350
15 dior oblique sneakers 13800
16 christian dior headband 13650
17 dior chuck taylor 11600
18 dior messenger bag 11200
19 dior clogs 9400
20 christian dior converse 9200
21 dior padlock necklace 8850
22 christian dior visor 8100
23 dior bifocals 7850
24 christian dior woven bracelet 7350

if q2.upper() == "YES":
    print('keyword = " ', kw_3,' " Top queries')
    print("")
    print(related_queries_dict[kw_2]["top"])

keyword = “ Chanel “ Top queries

if q2.upper() == "YES":
    print('keyword = " ', kw_3,' " Rising queries')
    print("")
    print(related_queries_dict[kw_2]["rising"])

keyword = “ Chanel “ Rising queries

Keyword suggestions

In this section, it is possible for the user, given the output of the previous analysis, to look for suggestions of new keywords or string related to a specific keyword or string.

This output can provide significant insights as it can reveal some words associations that companies might not be aware of.

keyword = input("Please write your keyword: ")

Please write your keyword: cartier watch

suggestions_dict = pytrend.suggestions(keyword)print(‘SUGGESTED RELATED KEYWORDS RELATED TO “‘, keyword, ‘“‘)
print(“”)
for number in range(0,len(suggestions_dict)):
    print(“\n”.join(“{}\t{}”.format(k, v) for k, v in suggestions_dict[number].items()))
    print("")

SUGGESTED RELATED KEYWORDS RELATED TO “ cartier watch “

mid /m/04799f2
title Cartier Tank
type Line of products

mid /g/11g02vjngy
title Sell Cartier Watch
type Watch store in London, England

mid /g/11dzswn26w
title The Watch Book
type Topic

mid /m/0gjkl
title Watch
type Topic

Additional output: Google Hot Trends

Lastly, is has been included this last section where the user can see what are the most popular searches performed in the last period and the searches of the current day.

Although not related with the keywords above, it can be useful to understand what are the key topics researched recently and to run specific marketing campaigns accordingly.

trending_searches_df = pytrend.trending_searches()
print(trending_searches_df.head())

0
0 Packers
1 Saints vs Bears
2 49ers
3 Giants
4 Buffalo Bills

Export trending searches into a csv file:

trending_searches_df.to_csv(“trending_searches.csv”, index=True)today_searches_df = pytrend.today_searches()
print(today_searches_df.head())

0 Packers
1 Saints vs Bears
2 49ers
3 David Johnson
4 Detroit Lions
Name: query, dtype: object

Export today searches into a csv file:

today_searches_df.to_csv("today_searches.csv", index=True, header=True)

Original files available here:

Google Trends — Gregorio Cioppa Betty Staccoli.ipynb