GA has an ability to feed its data to BQ, and from BQ it's much easier to get it to wherever. So it may be better for you to use GA API to stream the data through their measurement protocol (there are multiple libraries available for it). For the same reason, using your local DB may be a bad idea too. G Sheets won't be able to handle that amount of raw data. To that point, unaggregated data will take a lot of space. And you're unlikely to want that because that would be extremely annoying. If not, I would try doing my tracking on the front-end, too, but you mentioned that that's not an option, so I'm not gonna entertain the thought any further.įinally if those options aren't really options then you should think if you want to bother aggregating the data on the fly, in your script or not. Well, if your setup typically bypasses the web-server (which is annoying really), then the framework you use for managing endpoints would likely have semblance of logging in it. That would be the sturdiest and most reliable implementation and if done correctly, it won't even depend on your backend language, but just on the web-server logging. If what you've described is all you need, I would suggest feeding the access logs as they are to G sheets and be done with it. You have (or supposed to have) access logs on the backend. I always use its internal pseudo JS to move the data around. I never used Sheets APIs when the sheet is not opened. It will be annoying if your aim is just to get a few metrics and a few dimensions and never anything more. But later, when you want to get the data back, it's not that straightforward since those APIs are meant for pretty serious ETL process with exact understanding of all their internal definitions. They demand very certain data structure in order to be able to give fullest insights. You probably don't wanna complicate things with them. Note that it's extremely unlikely that you will be satisfied with just the device information in any analysis, but in this answer, I will presume that you know your situation better than I.įirebase/GA is large. Security side, now i store only User Agent string, but in future i could store more sensitive, so it's good idea to use this project to learn something reliable. The Cloud Function send record to Google Sheets, using some sort of Google Sheets Client SDK in Node.js.Īny good resource to understand Authentication in Google Ecosystem? When server receive a request, server send an other request to a Cloud Function hosted in Netlify or Similar.
#Newsflow install how to
Here i find difficult to understand how to setup Authentication, Client ID, API Key, etc in Google Console.(Never used before) When server receive a request, server send an other request to Google Sheets directly, using some sort of Google Sheets Client SDK in PHP. Here i find difficult to understand how to setup Authentication, Client ID, API Keys, etc in Google Console.(Never used before) When i want to analyze data i fetch the record from Firebase REST API in Google Sheets via Google App Scripts. When i want to analyze data i fetch the record from REST API in Google Sheets via Google App Scripts. Store records in a Wordpress custom table.Ĭreate a REST API inside Wordpress that read data from that table. I can use only the server to orchestrate all things. Įvery request the server receives is redirected to an external website. I don't have access to the browser's javascript. I'm asking here because maybe someone have been done this in the past and could tell his insights. I'm new to this and before start coding i'm doing research to define which "infrastructure" is better, and requires me less effort. I need also to be able to read records via Google Sheets.
Hence store a new record (containing the User Agent string) every time the server receive a request.
I want to study which devices send requests to a server (PHP - Wordpress).
0 kommentar(er)
