Hi everyone, we are working to release free access to a series of data prep nodes. We currently expose a series of AI and data prep algorithms through our Elipsa APIs but are looking to make it easier to use.
I am posting here to get feedback from the community on which (if any) data prep tasks would be useful to have wrapped into a single node:
Resample a collection of time series data
Fill missing information (options are fill with 0, specific value, forward fill, back fill)
Create lags in your time series data
There are other tasks that we are working on as well so any feedback on useful data prep and advanced analytics nodes would be helpful as we expand upon our existing Predictive Maintenance node that is now available in flows.
Those 3 bullets are certainly interesting and this topic is worth exploring further. I've long mentioned that I think an area somewhat lacking in Node-RED is the ability to more easily do ETL type tasks. Something that Python excels at but Python has nothing close to Node-RED
Thanks @TotallyInformation. That's the goal, basically take some of the functionality that can easily be done in python and wrap it in a node to combine it with the functionality of node-red. We also want it to serve a purpose and value for the node-red community as a whole so as we start creating them we will publish them to flows but if anyone has a 'special request' on ETL challenges definitely feel free to post to this topic.
The kinds of things I and people at work often need to do is work with large CSV documents. Combining data from different CSV's. Creating (sometimes reasonably complex) matching rules. Doing loads of data tidying from simply removing leading/trailing spaces and line feeds through to case manipulation and replacement of non-standard text.
I think that a good place to look is at the PowerQuery capability of Excel. The things you can fairly easily do there using point and click would be great tools in Node-RED as well.
There really aren't that many tools for doing this kind of work that are easy for non-programmers to use so introducing something via Node-RED would be a tremendous boon.
