Google Trends Category Search

[stochastiq]

Google trends has the option of choosing categories of search, rather than only terms. For example, I can choose to view "Shopping" category without entering any term to see the trend of all shopping related terms searched on google. How do I retrieve this series via the gtrendsR API? Furthermore, is there documentation for which number corresponds to which category on Google Trends?

[PMassicotte]

It is not supported at the moment, but it could be implemented simply I think.
https://www.google.com/trends/explore#cat=0-47-1190&date=1%2F2012%2012m&cmpt=q&tz=Etc%2FGMT-2

[stochastiq]

Hi Phillip,

Thank you for your reply.

When I choose the "Shopping" category (without including search terms) this is the url that pops up:
https://www.google.com/trends/explore#cat=0-18&q=&geo=SG&cmpt=q&tz=Etc%2FGMT-2

Correct me if I'm wrong, corresponding to this URL, the gtrendsR command looks something like this:
shop_trend<-gtrends(query="",geo=c("SG"), cat="0-18", start_date= "2004-01-01", end_date = "2016-04-05")

However, the command does not allow non-characters as a value for the "query" argument. So, I tried changing the url by adding a "&" to the query
https://www.google.com/trends/explore#cat=0-18&q=&&geo=SG&cmpt=q&tz=Etc%2FGMT-2

This url gives me the "Shopping" category series without any search terms. So the corresponding command I used is:
shop_trend<-gtrends(query="&",geo=c("SG"), cat="0-18", start_date= "2004-01-01", end_date = "2016-04-05")

However, the trends from shop_trend are different from that if I were to download the csv file from google trends directly. Firstly, the csv file contains figures in %, while the shop_trends trends are an index. Taking the percentage difference of the shop_trends trends also does not return the same figures as the csv file. Secondly, the csv file's top searches and shop_trends top searches differ, though they are in the same category.

Wonder if you would know how to implement the categories search for google trends using the existing gtrends command?

[PMassicotte]

If you are writing

shop_trend <- gtrends(query="&",geo=c("SG"), cat="0-18",
 start_date= "2004-01-01", end_date = "2016-04-05")

this means that you are searching the "&" character in the shopping category. I will try to implement this this week.

[stochastiq]

Ok, just as I suspected. I thought there would be a chance that it searches an empty string since the url
https://www.google.com/trends/explore#cat=0-18&q=&&geo=SG&cmpt=q&tz=Etc%2FGMT-2
shows the category search.

Anyway, thank you for your response, look forward to this extension :)

[PMassicotte]

Just started to implement this.

You can try it by installing the package with:

devtools::install_github("PMassicotte/gtrendsR", ref = "queries-on-categories")

> ch <- gconnect(usr, psw, verbose = TRUE)
Google login successful!
> data <- gtrends(geo=c("SG"), cat="0-18")
> head(data$trend)
       start        end shopping.SG
1 2004-01-04 2004-01-10           0
2 2004-01-11 2004-01-17           4
3 2004-01-18 2004-01-24           3
4 2004-01-25 2004-01-31          -1
5 2004-02-01 2004-02-07          -1
6 2004-02-08 2004-02-14           0
> plot(data)
> data <- gtrends(geo=c("CA"), cat="0-20-263", res = "1d")
> head(data$trend)
                start sporting.goods.CA
1 2016-04-06 14:00:00                 0
2 2016-04-06 14:08:00               -12
3 2016-04-06 14:16:00                 6
4 2016-04-06 14:24:00                 1
5 2016-04-06 14:32:00                -6
6 2016-04-06 14:40:00                -3
> plot(data)

[stochastiq]

Hi Phillip,

Thanks for getting back to me. It appears that I must not be doing
something right, because I still get the error:
Error in names(trend) <- make.names(paste(kw, geo)) :
'names' attribute [1] must be the same length as the vector [0]

I reinstalled the packages, and used the code:

library(gtrendsR)
user <- "XXXXX"
psw <- "XXXXX"
gconnect(user, psw, verbose = TRUE)
shop_trend<-gtrends(query="",geo=c("SG"), cat="0-18", start_date=
"2004-01-01", end_date = "2016-04-05")

Referring to your latest reply, may I ask what do you mean by:
"You can try it by installing the package with:

`i"
When I run your code above, I get this error:
Error in stopifnot(is.character(query), is.vector(query), length(query) <=
:
argument "query" is missing, with no default

Thank you for your help.

On Thu, Apr 7, 2016 at 9:59 PM, Philippe Massicotte <
notifications@github.com> wrote:

Just started to implement this.

You can try it by installing the package with:

`i

ch <- gconnect(usr, psw, verbose = TRUE)Google login successful!> data <- gtrends(geo=c("SG"), cat="0-18")> head(data$trend)
start end shopping.SG1 2004-01-04 2004-01-10 02 2004-01-11 2004-01-17 43 2004-01-18 2004-01-24 34 2004-01-25 2004-01-31 -15 2004-02-01 2004-02-07 -16 2004-02-08 2004-02-14 0> plot(data)> data <- gtrends(query = "", geo=c("CA"), cat="0-20-263", res = "1d")> head(data$trend)
start sporting.goods.CA1 2016-04-06 14:00:00 02 2016-04-06 14:08:00 -123 2016-04-06 14:16:00 64 2016-04-06 14:24:00 15 2016-04-06 14:32:00 -66 2016-04-06 14:40:00 -3> plot(data)

[image: rplot]
https://cloud.githubusercontent.com/assets/4519221/14353494/99b06966-fcd9-11e5-8a40-db276f673089.png

[image: rplot01]
https://cloud.githubusercontent.com/assets/4519221/14353499/9d0021b0-fcd9-11e5-871f-7471ee5051a6.png

—
You are receiving this because you authored the thread.
Reply to this email directly or view it on GitHub
#89 (comment)

[PMassicotte]

You have to install this specific branch if you want to try it. You have to type it in R.

devtools::install_github("PMassicotte/gtrendsR", ref = "queries-on-categories")

[dap2rb]

Hi PMassicotte,

thank you for your package. Could you clarify two issues:

1. How to install the branch. I type: "devtools::..." in R and it gives me the following response: "Error in loadNamespace(name) : there is no package called ‘devtools’"
2. How do I download items for the cities. I tried maimi with geo=US-FL-528 and it did not work

thank you a lot

[eddelbuettel]

For Pete's sake please don't hijack random other threads. Please delete the message above, @dap2rb .

[BERENZ]

Dear All,

my students (@KasiaaZ and @MrKrzysztof) prepared list of categories that are used in Google Trends (included in MS Excel file google_trends_categories.xlsx).

Maybe you would like to include this list into package vignette?

Best, Maciej

google_trends_categories.xlsx

[PMassicotte]

Hi @BERENZ

That sounds like a good idea. How did you retrieved it? I downloaded the json category file from Google a while back, but ran out of time to process it.

[BERENZ]

Hi @PMassicotte

It wasn't automated...I think they just used a copy-paste method.:)

Do you have a link to download json file? I could write a script to parse the categories.

[PMassicotte]

@BERENZ

categories.json.zip

You can read them using:

categories <- jsonlite::fromJSON(txt = "data/categories.json", flatten = TRUE)

[BERENZ]

@PMassicotte thanks! I will provide do it tommorrow.

Where did you find this file? Is it somewhere online?

[PMassicotte]

@BERENZ I extracted it from my Chrome browser using the dev console while browsing Google Trends web site. CTRL + SHIFT + C

[jayruffell]

Hi, is it still possible to get search volumes for a given category without providing a keyword? The branch that this feature was on devtools::install_github("PMassicotte/gtrendsR", ref = "queries-on-categories") doesn't seem to exist anymore. Thanks

[PMassicotte]

I think you can do it with the current version.

[jayruffell]

Sorry yes should've checked the current version - code works thanks