Learn by Directing AI
All materials

data-dictionary.md

Data Dictionary: BrightSmile Dental Booking Data

Column Definitions

Column Type Description
date DATE (YYYY-MM-DD) Date of the booking
clinic_id TEXT Clinic identifier. One of: Jinniu, Wuhou, Chenghua, Qingyang, Shuangliu, Gaoxin
patient_type TEXT Either "new" (first visit at any BrightSmile clinic) or "returning" (has a prior visit)
service_category TEXT One of: general, cosmetic, orthodontics, implant
booking_source TEXT How the patient was acquired. One of: walk-in, online_booking, referral, wechat_ad, kol. May be null for walk-in patients where the source was not recorded
revenue DECIMAL Revenue from this booking in Chinese Yuan (CNY). Always positive

Business Terminology

  • New patient -- A patient visiting any BrightSmile clinic for the first time. A patient who visits Jinniu first and then Gaoxin later is "returning" at Gaoxin
  • Returning patient -- A patient who has visited any BrightSmile clinic before this booking
  • Booking source -- The acquisition channel that brought this patient in:
    • walk-in -- Patient walked in without a specific referral or ad exposure
    • online_booking -- Patient booked through the BrightSmile website
    • referral -- Patient was referred by an existing patient. Before the campaign, all referrals were organic. During the campaign, referrals include both organic and bonus-driven (CNY 200 referral bonus program). The data does not distinguish between these before the campaign start date
    • wechat_ad -- Patient booked after seeing a WeChat advertisement. Only tagged if the patient booked through the WeChat link. Patients who saw the ad but booked by phone are not tagged (estimated 15-20% uncaptured)
    • kol -- Patient booked through a local Key Opinion Leader (KOL) partnership link or code
  • Service categories:
    • general -- Routine checkups, cleanings, fillings, extractions
    • cosmetic -- Teeth whitening, veneers, bonding
    • orthodontics -- Braces, aligners, retainers
    • implant -- Dental implant procedures

Notes

  • The dataset covers two full calendar years of bookings across all six clinics
  • The Gaoxin clinic opened on October 15 of the second year. All records for Gaoxin begin on that date. Because it is a new clinic, all patients there are classified as "new" -- this is factually correct but can inflate new-patient metrics if Gaoxin is included in campaign analysis
  • Bookings typically increase in Q4 (October-December) due to scheduling before Chinese New Year and school holidays. This seasonal pattern exists in both years, independent of any marketing campaign
  • The marketing campaign ran from October 1 to December 31 of the second year
  • booking_source has approximately 3% null values, primarily from walk-in patients where the front desk did not record the source