Who are the Unhappy Buddies
Competition Goal: Forecast the churn_dummy field
Train Data Set • 50 K train customers data, with churn_dummy variable • Transactional table related to 50 K train customers data 2. Live Data Set may be used as additional Train Data Set • 50 K application customers data, where churn dummy variable may be derived • Transactional table related to 50 K application customers data 3. Application Data Set • 50 K application customers data, without churn_dummy variable • Transactional table related to 50 K application customers data
About Datasets:
Some fields related to customers characteristics, that are • id : randomly generated • gender • tenure_days : number of days since the subject is a buddy • age • primary _address_city : the city (with * we do refer to «other», the cities with only a few of • some dummy variables: respectively being or not an italian citizen, a student, a worker, have a promotion, own the love profile, have closed the acount (churn) • estinction_date : account closing date, evaluated only for churn=1
For each customer, we do have a number of transactions (can be also zero) • ids_id : the foreign key for linking to the customers table • tp_mov_gk : transaction tipe • co_naz_iso : country where the transaction took place • dv_mov : currency • im_mov : transaction amount • tipo_carta : card type • new_dt_ope : day and hour of the transaction