Improvement: use getInsights() for longer periods of time

[daroczig]

Improvement: use getInsights() for longer periods of time

[daroczig]

You mean the fb_insights function? As this can do async queries as well, there shouldn't be a limit on the amount of data to be pulled from the Marketing API. How many ads and what time period are you pulling?

[tokumotion]

@daroczig I am pulling 30ish campaigns for the lapse of a year. Each campaign should have around 3 ads each.

[daroczig]

Can you post here a demo function call when you hit the limit and the response from FB? Of course please sanitize the account id, token etc.

[tokumotion]

Hi @daroczig, this is my example.

fblista <- fbad_list_campaign(fields = c("id", "name", "objective", 
                                         "start_time","stop_time"))

fblista %<>%
  filter(grepl("brand1|brand2|brand3", fblista$name, 
               ignore.case = TRUE),
         start_time >= "2015-01-01", stop_time <= "2015-12-31")

fblista[,c(4:5)] <- lapply(fblista[,c(4:5)], ymd_hms)

POST_ENGAGEMENT <- filter(fblista, objective == "POST_ENGAGEMENT")

How does POST_ENGAGEMENT look like?

> dim(POST_ENGAGEMENT)
[1] 56  5

This is where I find some troubles, if I go with job_type = "async":

> fb_insights(target = POST_ENGAGEMENT$id,
 +             fields = c("impressions", "social_impressions", "clicks", 
 +                        "social_clicks", "spend", "website_clicks",
 +                        "reach", "ctr", "unique_ctr", "cpc", "cpm",
 +                        "cpp"), job_type = c("async"))

I get this error:

 Error in fb_insights(target = POST_ENGAGEMENT$id, fields = c("impressions",  : 
  Batched queries are not possible with async call. Please query only one item at a time.

If I go with job_type=c("sync")

> test <– fb_insights(target = POST_ENGAGEMENT$id,
 +             fields = c("impressions", "social_impressions", "clicks", 
 +                        "social_clicks", "spend", "website_clicks",
 +                        "reach", "ctr", "unique_ctr", "cpc", "cpm",
 +                        "cpp"), job_type = c("sync"))

I don't get all the data I made a query for (dim() should be returning 56 rows):

> dim(test$`0`)
[1]  4 14

[daroczig]

What about querying all your campaigns in your FB account for a time period (using eg date_preset or date ranges) instead of passing a list of campaign ids (by omitting the target parameter)? That should work with the async methods as well.

Another option could be doing separate queries, something like:

lapply(POST_ENGAGEMENT$id,
       function(id)
       fb_insights(target = id,
                   fields = c("impressions", "social_impressions", "clicks",
                              "social_clicks", "spend", "website_clicks",
                              "reach", "ctr", "unique_ctr", "cpc", "cpm",
                              "cpp"), job_type = "async"))

Then calling rbindlist from data.table or similar on the results.

[tokumotion]

I'm getting this:

test_2 <- lapply(POST_ENGAGEMENT$id,
   function(id) fb_insights(target = id,
                 fields = toJSON(c("impressions", "social_impressions", 
                            "clicks",
                            "social_clicks", "spend", "website_clicks",
                            "reach", "ctr", "unique_ctr", "cpc", "cpm",
                            "cpp"), job_type = "async")))

This is returns a list with each campaign's insights.

Note that the original code would return error if you don't convert the fields toJSON().

[daroczig]

So it works, @tokumotion? If you have the list, you should be able to merge those together via rbindlist from data.table or do.call and rbind, right?

[tokumotion]

@daroczig Yes, it works. I'm currently working on a way to convert the results into a data frame. Thank you!

This code converts the outcome into a dataframe:

test_2 <- do.call("rbind", lapply(test_2, data.frame))

There is something I don't quite get, if you see the code I placed, the toJSON() function encloses the job_type = "async"parameter. Running this code will return the desired outcome. However, if you let job_type="async" out (as one would intuitively would), the code returns this error:

> test_2 <- lapply(POST_ENGAGEMENT$id,
+        function(id) fb_insights(target = id, 
+                                 fields = toJSON(c("impressions", 
+                                                   "social_impressions", 
+                                                   "clicks", "social_clicks", 
+                                                   "spend", "website_clicks",
+                                                   "reach", "ctr", 
+                                                   "unique_ctr", "cpc", 
+                                                   "cpm", "cpp")), 
+                                 job_type = "async"))

Error:

Error: Argument 'txt' must be a JSON string, URL or file.
Called from: base::stop(..., call. = FALSE)

Why would you have to insert job_type="async"into the JSON file?

[daroczig]

Sounds good, thanks for the feedback -- I am closing this ticket. Can you please open a new ticket for the job_type issue with a minimal example?