Automate Testimonials in Strapi with n8n

hey i'm tom and i quickly wanted to show you how this page works um as you might have seen this is like a quick example page that contains a bunch of posts that mostly talk positively about ntn or strappy and if you click on new you can see like a little form pops up where you can add your own post um but there are also a bunch of tweets that you could open by clicking on them now you're probably familiar with scrappy since you joined strappyconf and there's really nothing exciting to the structure of this website posts are just their own content type and if you open one of these you can see they have obviously the content an author the url and the created date and that's about it as far as strap is concerned it works great this page fetches the post from the api straight away like whenever you load it it's live data so it's blazing fast really fun to use but what's more interesting here is how these posts are created and managed because there is no manual work involved like you don't have to go here click on add new posts this all happens automatically through n8n and we'll take a look at um both cases like first the form case whenever you fill out the form that i've shown you just here like this little thing what happens is the uh system tally that hosts this form since webhook to n so we look at this web hook we can look at all the data and it and has received from telly here um obviously lots of uh http headers not that interesting but there's also the text that uh i have entered in the form and my name obviously this is then processed by n in the first step i'm just simplifying it a bit right so we need the content author create a date url everything else i throw away um and i'm then sending it to google for sentiment analysis so if we open this node here uh you can see it does have a bunch of settings on the left they currently grayed out because we're looking at past data you can't manipulate past data um but you can still see what's um happening here right so there's a set of credentials that's been selected here um i'm using the analyst analyze sentiment option um and the content obviously is in it and it's awesome and so is strappy on the right side we can see the response coming back from google um google assigns a score to all these sentiment analysis um score below zero means it has a negative sentiment score above zero up to one means a positive sentiment you can see this is a very positive phrase and has gotten a score of 0. 9 from google so just a short of one on the next step i'm just merging the original input data and the output data from google back into one object so what we have here is the original content as well as the analysis from google i'm then checking if the score is very positive so if it's above 0. 4 and if so the item is sent to stroppy so and it n has a pre-built integration for straw p this drop you note if you look into this it looks very similar to the google uh natural language note that you've seen earlier right so it just has the credentials here and i'm then saying create a new entry of type posts with these columns that um it simply takes from the previous node so these four here um and then it gets added to strappy uh simple as that and that happens automatically right once this has been set up obviously no further user interaction is required um quickly take a look at this tom droppy still great still going strong let's say something like that see positive enough but it should be right so if you reload the page now you can see the new post has appeared here and um it's obviously also in straw piece if you reload the post page here we go this is our new post um now the other case are these tweets you can see like the last three messages here are all these uh manual posts and everything below here is tweets right you can click on them to open them if needed now the reason uh they do not come in real time is because this shows a polling approach right so in this case no data gets actively centered in it then instead i'm simply running this workflow every 30 minutes this can of course be configured right you can add whatever you want in here um and then going to twitter and simply search for tweets containing strappy or n in the english language and then it's pretty much the same from here i'm first simplifying the result again just looking at content author created url i don't care about the rest um i'll do one more step of filtering here compared to this is where this workflow is a bit more complex i check

if the tweet is a retweet on older tweet don't care about these lots of retweets are boring so i only want the original tweets so um i'm checking here is does the content start with rt at which is how twitter still marks retweets internally and um is it a relatively recent tweet if so the process is exactly the same data gets passed on to google for sentiment analysis and then merging it with the input data again and then validate does it have a positive sentiment yes or no if it does have a positive sentiment it makes its way onto the page just like that again three times the charm here from the looks of it and all these three posts appear on this page here and that's how nhn can automate content ingestion in strappy you