Hello fellow redditors,I'm working on an IoT project for school and one of the most important parts we need to make is Data Analysis. To be more clear, I am getting temperature, pressure, current usage from a few sensors and send them to my server to be handled through MQTT.All is fine and dandy until some more advanced analysis things (at least for me). Right now we have some time-series data we can use on the frontend, but there are a couple of things we want to implement for a better grade.One thing would be: if the current usage changes, the app will notify the user and ask what changed the usage (maybe I plugged my laptop out and put my phone charger). We also want to make some predictions for temperature change in the next days.What I want to know is, what would be the best way to find out when to send a notification to the user? Like, the current readings are sometimes wrong so I don't want to send it instantly, I want for it to wait a bit so those errors are ignored. How should I optimize the predictions? (we are using a regression algorithm at the moment that gave us some results, but how can I optimize the algorithm to not do those calculations all the time?) I want to mention that we want some real-time data on the frontend (that kinda works for the basic time-series data through sockets)Hope I was clear enough.I though of using Bull for queuing the data analysis part but I never used it. Any other ideas?
Submitted October 13, 2019 at 02:32PM by razzzey
No comments:
Post a Comment