Re: Creating groups based on record values

857 1
Showing results for 
Search instead for 
Did you mean: 
5 - Automation Enthusiast
5 - Automation Enthusiast

Sorry, this is quite a convoluted ask and I understand it may be too complex for a simple response.

Context: We have a community of 60 individuals recorded in an airtable. We run a limited dinner series to allow them to network with as many people as possible in as few instances (say 4 - 6 only). We manually generate groups of five persons for each dinner using various criteria such as persona, interest, gender, etc. In some instances, we wish to increase the chances of similar profiles meet. In others, we want to decrease chances.

Ask: Is there a way to use automation or script in Airtable to automate this group generation? We would also like to ensure it creates a value in response to “have these individuals been grouped before” so we can increase or decrease that criteria.

4 Replies 4

Hi Darren, with a script this is definitely doable. Automations might work too, depending on how the matching works

If you could provide more details on how people are selected and stuff that’d be helpful

5 - Automation Enthusiast
5 - Automation Enthusiast

Amazing! Thanks, Adam.

There’s quite a dense network of properties/criteria. We may not use all properties in every query, but it should “learn” who has been matched together previously. Each is attributed a weight for (10) or against (-10) to generate random groups.

  • spikeType
  • startupIndustryFocus
  • industryExperience
  • birthCountry
  • nationalities
  • functionalExperience
  • gender
  • workedTogether (negative value would group people that have previously been together)
  • preferences (their preferences to others, 0-5)

If we can make this work, another win would be being able to add new properties as they develop.

Ah okay, then I’d probably recommend doing it via a script then

Off the top of my head, for the one where you’d want to find people that match well, I think I’d approach that by picking out 12 random people each time and structuring the group around that one individual? That seems like the simplest way to apply the weightages

Thinking about getting a script out that would do exactly what you needed to do made my head hurt heh

Ha, god, i can relate to the headache. Doing this manually is torture.

Thanks for the suggestion!