In [ ]:

    

import dir_constants as dc
from tqdm import tqdm_notebook


    

In [ ]:

    

platform = 'lendingclub'

store = pd.HDFStore(
    dc.home_path+'/justin_tinkering/data_science/lendingclub/{0}_store.h5'.
    format(platform),
    append=True)


    

In [ ]:

    

# some setups for code block below
pmt_hist_ids = store['pmt_hist_ids'].astype(int)
max_id = pmt_hist_ids.max()
chunksize = 800
n_chunks = len(pmt_hist_ids)//chunksize + 1

interesting_cols_over_time = [
    'outs_princp_beg',
    'all_cash_to_inv',
    'date',
    'fico_last',
    'm_on_books',
    'status_period_end',
    'loan_id',
    'loan_id_num',
]

disc_rates = np.arange(.05,.36,.01)


    

In [ ]:

    

printed_cols = set()
npv_roi_holder = {}
for n in tqdm_notebook(np.arange(n_chunks)):
    if n == 0:
        left_bound = 0
    else:
        left_bound = pmt_hist_ids[n*chunksize]
    if n == (n_chunks - 1):
        right_bound = max_id
    else:
        right_bound = pmt_hist_ids[(n+1)*chunksize]
    
    chunk = pd.read_hdf(
        store,
        'pmt_hist_intermediary_2',
        where='(loan_id_num > left_bound) & (loan_id_num <= right_bound)',
        columns = interesting_cols_over_time)
    
    id_grouped = chunk.groupby('loan_id')
    for ids, group in id_grouped:
        npv_roi_dict = {}
        funded = group.iat[0,0]
        cfs = [-funded] + group['all_cash_to_inv'].tolist()
        for rate in disc_rates:
            npv_roi_dict[rate] = np.npv(rate/12, cfs)/funded
        npv_roi_holder[ids] = npv_roi_dict


    

In [ ]:

    

npv_roi_df = pd.DataFrame(npv_roi_holder).T
npv_roi_df.columns = npv_roi_df.columns.values.round(2)


    

In [ ]:

    

store['loan_npv_rois'] = npv_roi_df
store.close()


    

In [ ]:

    




    

In [ ]: