Table For Two

Table for Two is a web app that allows Mixpanelers the chance to meet with other Mixpanelers from different departments whenever they'd like. It prioritizes new hires, and allows lunch and coffee dates in 30 minute increments.

Features

• Set your own recurring availability
• Set your own profile settings (location? google hangout?)
• Matches folks automatically and puts the event on the calendar

How to run on your machine if you have everything installed

• activate an env with source tablefor2/bin/activate
• run with python manage.py runserver
• go to localhost:8000

How to run the matching system

• Go to Heroku
• Run python manage.py match_users in the console
• Copy the output of the URL into the console

V2

• • Exclude to Mixpanel addresses --> non-Mixpanelers can sign up but not be matched
• Programmatic frequencies
• Cron job
• Move to AWS?
• Variable locations?
• See your Mixpanel calendar?
• Show EMEA/NYC/SF friendly days?

V1

Signup

• Create database and model
• Create signup flow
• Create index-logged-out page
• Ask for extra information (edit-profile)
• Create the form and edit-profile page
• Save user in database
• Create profile page
• Show existing profile in form

Availability / Recurring Availability

• Create front-end for availabilities
• Save recurring availabilities in the back-end
• Edit availabilities
• Test
• [] Use UltiPro to prevent scheduling when a user is OOO
• [] Use UltiPro to turn off matching for a user if they've left Mixpanel

Matching

• Match the people by availability and 1x/wk (frequency V2)
• Show your future matches
• Show your previous matches
• Send a google calendar invite
• Change everything to proper timezone
• Option to turn off matching if you want to stop or you're OOO
• Refactor to once a month
• Change to new
• Test
• Create a cron job

Mixpanel

• Goal: acquisition
• Signup flow (index -> signup -> profile -> save)
• identify
• Save recurring availability --> double, because 2 POST requests
• Match made / invite sent (server-side)
• People prop: email, dept, etc, number of matches, number of availabilities
• Test

Front-end

• PRETTIFY
• index-logged-out
• index-logged-in
• edit-profile
• edit-availability
• make sure each display and browser is good
• tool tips to onboard users about things
• connect to heroku properly

Matching process

• For every recurring availability, create an Availability for that user (date and time of beginning 1/2 hour, assuming timeslot is half hour)
• Runs a thing at 3pm the day before
• Organize all the users based on their hire date
• Match the newest employees with the most senior employees
• For each pairing, check for:
  • If a user is currently accepting matches
  • Only if fits their frequency of 1x/wk (V2 will be programmatic)
  • People who are available at the same time
  • Are in a different department
  • Haven't matched before
  • Those who are in the same location (last, GHangout)
  • User can choose between veterans or new hires
• If that user is matched, then we'll set the name and email equal to the match
• If the user is not matched, keep going to the next least senior employee
• If someone does not have a match continuing this process, check the next person

Mixpanel Implementation

Events

• Page Viewed
  • Index-logged-out (home page)
  • Index-logged-in (dashboard)
  • Edit-availability (edit availability)
  • Edit-profile (edit profile)
• Profile Created
• Recurring Availability Saved
• Match Created
• Calendar Invite Sent

Properties

• Page Viewed
  • Page
• Profile Created
• Recurring Availability Saved
• Match Created (2)
  • Current User Department
  • Current User Location
  • Other User Department
  • Other User Location
  • Google Hangout or In Person
• Calendar Invite Sent (2)
  • Time
  • Timezone
• People
  • $first_name
  • $last_name
  • Preferred Name
  • $email
  • Department
  • Accepting Matches
  • Google Hangout
  • Location
  • Timezone
  • Frequency
  • Date Entered Mixpanel
  • Number of Matches
  • Number of Recurring Availabilities
  • Last Match Created
  • Date Joined

Other Implementation Details

• saving of distinct_id
• identify on db's distinct_id
• clear cookie upon logout

Sample data

Hermione Granger

Success

June 19, 2019

Ron Weasley

Support

April 10, 2019

You have an upcoming match with Harry Potter on July 31, 2019, noon via video call! They are from Engineering. A fun fact is that "I know how to fly a broomstick". Their favorite movie is Harry Potter and the Sorcerer's Stone.