Time series data storage strategy and visualisation in Node Red

Dashboard
1

Are you aware of an existing node that could be used for visualising time series data?

I would like to analyse local electricity consumption, per day, per week, per month. Depending on the level of resolution, the data should be aggregated and visualised either per hour, per week or per month.

The consumed power is measured by a Carlo Gavazzi power meter. Currently, the cumulative power reading is read over Modbus and stored in a PostgreSQL table once per minute (the cumulative power reading is a stock variable). In addition to stock variables there are flow variables, e.g. temperature, and heat pump compressor modulation (heat pump compressor modulation affects cumulative power consumption)

Stock variables are easy to visualise, because their value at a given time should be know. Flow variables are more tricky to visualise. What should the value for a flow variable be for a given month? As I see it, flow variables would be recorded continuously, and averaged at the time of visualisation, e.g. temperature over a given month.

2

The chart node?

But I am not sure that will do all you need.

3

I use Grafana for similar application

1 Like
4

While you can throw that kind of data at a Dashboard chart, you generally don't get much flexibility for doing aggregations.

So most of us dump timeseries data into InfluxDB which is specifically designed to handle such data and provides ease methods to aggregate by time.

Once you go down that route, the use of Grafana - which has good integration with InfluxDB - makes sense since it has point-and-click type tooling for building charts and dashboards from the data and has built-in variables that let you change time-spans and time aggregations.

Typically SQL db's such as Postgres don't handle time series data well.

5

This topic was automatically closed 30 days after the last reply. New replies are no longer allowed.