MoneyClassifier.py

from textblob.classifiers import DecisionTreeClassifier as TextClassifier
from datetime import datetime
import matplotlib.pyplot as plt
import csv

num_months = 13

with open('train.csv', 'r') as fp:
    
    clf = TextClassifier(fp, format="csv")

activity = {}
months = []

with open('ChaseActivity.csv', newline='') as csvfile:
    
    reader = csv.DictReader(csvfile)
     
    for row in reader:
         
        date = datetime.strptime(row['Posting Date'], '%m/%d/%Y')
        desc = row['Description']
        amt = row['Amount']

        month = date.strftime('%b %Y')
        label = clf.classify(desc)

        if month not in months:
            months.append(month)

        if not label in activity:
            activity[label] = [0] * num_months

        if not month in activity[label]:
            activity[label][months.index(month)] = 0

        activity[label][months.index(month)] += -float(amt)

        print(desc)
        print(label)
        print()

assert len(months) == num_months

labels = clf.labels()
labels.remove('Misc')
labels.remove('Investments')
labels.remove('Transfer')

def flip(list_): # B/c CSV was backwards
    return list(reversed(list_))

def _range(n):
    return list(range(num_months))

plt.stackplot(_range(num_months), [ flip(activity[label]) for label in labels ])
plt.legend(labels)
plt.ylabel('$')
plt.xlabel('Month')
plt.xticks(_range(num_months), flip(months))
plt.show()

#print(clf.pseudocode())