Facebook PixelHow to record live social interactions as data in the easiest possible way?
Create newCreate new

How to record live social interactions as data in the easiest possible way?

Image credit: https://www.pexels.com/photo/two-men-standing-near-wall-2247072/

jnikola Nov 28, 2022
Please leave the feedback on this challenge

Is the problem still unsolved?


Is it concisely described?

Bounty for the best solution

Provide a bounty for the best solution

Bounties attract serious brainpower to the challenge.

Currency *
Who gets the Bounty *
Figure out how to record every live interaction in the easiest possible way. The idea should aim for the solution with the minimum possible effort and the best data accuracy.
How I came to this challenge?
I have a constant problem with not remembering the names or even the faces of people I meet. I often find myself talking to someone and trying to figure out why the person greets me in a friendly way as we have already met. Usually, after several hours, I remember that I spent a day with that person at a conference and we had a nice talk. Or I don't remember at all. Because of that, I searched for tools that could help me to stay organized and memorize people and names. I found Customer Relationship Management apps (described in more detail here) that store all your contacts from all social media accounts, mails and contact lists on a one place, track and update changes in the contact information and record virtual interactions with them. While private messages on Facebook are easily tracked, these apps lack the real-world interaction recording system that is easy and efficient. Some people offered their "easy" solutions through spreadsheets, Google Forms, Airtable or similar software that offer a bit easier manual input of the information. However, ...
  • ... there is no easy way to detect and record/keep track of real-world interaction, besides manual input
  • tracking information about your social interactions could help you:
  • remember names and topics of your previous conversations with someone
  • get statistics on how many times you saw someone
  • get reminded to contact someone who you didn't contact for a long time
  • find contacts that match your current interests or can solve current problems (filtering by education, company, skills, met location, topics of interest, itd)
  • finding more compatible party invitees -> better parties
  • find perfect gifts for someone by additional tools
  • it's a feature that could be used on business meetings, conferences and other professional events to easily find and connect
Some questions to start the ideas going
  • How could we detect someone's presence around us?
  • How could we distinguish between the presence and actual social interaction (sitting in a bus or at the conference vs talking to someone)?
  • How could we record this meaningful interaction with high precision?
Creative contributions

Using smartwatches and other devices to detect social interactions

Subash Chapagain
Subash Chapagain Nov 30, 2022
I was wondering if we could use existing smart devices, phones and similar equipments to record the frequency and length of various social interactions we might have during the day. For instance, the phone tracks whenever we speak, and it detects the change in pitch/intensity of the voice to determine if we are conversing with other people. Machine learning for language detection could be helpful here. I think this will be useful in answering your second question in the last section. The actual physical input (intensity, pitch) of the human voice combined with the machine learning data that differentiates whether the conversation is bidirectional versus segmented, discrete noise could be one model. Regarding precision, like all machine learning algorithms, it should get better and better with time as more and more data accumulates. Similarly, devices like smartwatches could be developed to have a feature that records the change in heart rate or blood sugar levels whenever we are engaged in a social interaction.
Also, use smartglasses for visual+audio data to register actual social interactions. All of these combined should offer at least a starting point for recording social interactions.
Please leave the feedback on this idea
jnikolaa year ago
Great ideas! I was on the same track of thoughts. I like the recognition of the conversations which you proposed. I think it could work with good accuracy. Maybe the biggest problem could be recognition of "unimportant" conversations, for example, with the lady at the cashier, or in the bank.
Please leave the feedback on this idea
Subash Chapagain
Subash Chapagaina year ago
J. Nikola If you go by the neurological evidence for what actually counts as a 'social interaction'; there is no direct way to say for which interaction is important and which is not. Andrew Huberman, the stanford Professor and neurologist makes this point very clear in several of his podcast episodes: any kind of social interaction is a good social interaction. For an interaction to benefit, it always need not involve deep diving conversations about life, philosophy and love. Any thing as trivial as routine greeting and weather-talk is beneficial. It is just not the depth of the conversation, rather the very engagement with other humans, to vocalise and to pay attention to what they are saying that makes the positive impact. Combined with other cues like eye-eye interaction and touch, these positive effects are compounded.
Please leave the feedback on this idea
jnikolaa year ago
Subash Chapagain Hehehe I agree with you, but what I wanted to refer to as "unimportant" was an unintended or spontaneous interaction without information sharing or a purpose (a.k.a. small talk with the cashier lady). I completely agree with you in the psychological sense, but I think that for the "social interaction" log, where you want to know the name and time of the person you talked to, these interactions are not important. Now when I wrote this I came to an idea.
Name sharing in the introduction as a trigger to record the interaction
Could every interaction with a person of an undefined identity (signal, voice or location not recognised before) be distinguished from unimportant ones by the simple name exchange? If we meet and we tell each other names, does it mean that interaction is more important than the talk with a cashier? In other words, does introducing your name to a cashier make your interaction with a cashier more important? I think yes :)
Please leave the feedback on this idea

Detecting the signal of the person you are interacting with

jnikola Dec 01, 2022
Since we always have some tech in our pockets or on our wrists, it would be cool to somehow use the signal of this tech to detect WHO is the person we are interacting with.
How would it work?
I imagine it to work based on Bluetooth or a similar wireless connection. When you approach the person and start a conversation, the algorithm defines it as a social interaction between you and someone and starts to search for signals around you. If the signal is new, it gets remembered and the algorithm asks you to fill in the information of the new contact. If the person is in your contacts and you have already interacted with it, the signal of the device is detected and the app asks you "if it was...", for example, "...John that you had a conversation with at 12:03 am?". You confirm later when you have time or you see the notification and it gets registered as an interaction.
Now, I am not a tech guy and I am not sure how could this be done in reality. Bluetooth is pretty simple example, but it should be always turned on on all smartphones or other devices of people you are interacting with. I am not sure if there is a way to do the same thing via network connection or other radiowaves.
Additional features
Close-by friends alert
If the signal of a person from your contact lists is well known and you defined a person to be on your "close friends" list, than you could activate a feature that alerts you when John is again close to you and you didn't interact already with him for a minute or two.
Please leave the feedback on this idea

Add your creative contribution

0 / 200

Added via the text editor

Sign up or


Guest sign up

* Indicates a required field

By using this platform you agree to our terms of service and privacy policy.

General comments

Shubhankar Kulkarni
Shubhankar Kulkarnia year ago
The black mirror episode of "The entire history of you" came to mind :)
Please leave the feedback on this idea