Change net income figure; Add table

README.md

@@ -1,3 +1,22 @@
 # banking-breakdown
 
-Visualize banking statements.
+Visualize bank statements.
+
+## Usage
+
+1. Assign categories to bank statement entries:
+    ```bash
+    $ python -m banking_breakdown categorize -i [bank_statement.csv] -f [regex_file] 
+    ```
+2. Generate report from categorized data:
+   ```bash
+   $ python -m banking_breakdown report -i [categorized.csv]
+   ```
+
+## Other
+
+### Generate GUI from `*.ui` file
+
+```bash
+$ pyuic6 res/main_window.ui -o banking_breakdown/ui/main_window.py
+```

banking_breakdown/__main__.py

@@ -1,35 +1,61 @@
-import pandas as pd
-from banking_breakdown import document_builder
-import subprocess
-import os
-import shutil
+from banking_breakdown import ui, regex_categorizer, statement_parser, \
+    document_builder
 
-import pandas as pd
+import argparse
 
-from banking_breakdown import types
-import numpy as np
+
+def categorize_func(args):
+    import pandas as pd
+
+    df = pd.read_csv(args.i, delimiter=args.d)
+
+    if args.f is not None:
+        df = regex_categorizer.assign_categories(df, args.f)
+
+    import signal
+    signal.signal(signal.SIGINT, signal.SIG_DFL)
+
+    ui.show_main_window(df=df)
+
+
+def report_func(args):
+    report_data = statement_parser.parse_statement(args.i)
+    document_builder.build_document(report_data)
+
+
+#
+# Define CLI
+#
 
 
 def main():
-    categories = ["A", "B", "C", "D", "E", "F", "G"]
-    values = np.array([10, 12, 53, 12, 90, 23, 32])
-    values = values / values.sum() * 100
+    parser = argparse.ArgumentParser(prog='banking_breakdown',
+                                     description='Visualize bank statements')
 
-    total_value = np.random.normal(size=10) + 4
-    net_income = np.diff(total_value)
+    subparsers = parser.add_subparsers()
 
-    category_overview_df = pd.DataFrame(
-        {"category": categories, "value": values.astype('int32')})
-    t = np.linspace(0, total_value.size, total_value.size)
-    total_value_df = pd.DataFrame({"t": t, "value": total_value})
-    t = np.linspace(0, net_income.size, net_income.size)
-    net_income_df = pd.DataFrame({"t": t, "value": net_income})
+    categorize_parser = subparsers.add_parser("categorize")
+    categorize_parser.set_defaults(func=categorize_func)
+    categorize_parser.add_argument('-i', required=True,
+                                   help="Bank statement CSV")
+    categorize_parser.add_argument('-f', required=False,
+                                   help="JSON file containing regexes to"
+                                        " pre-categorize statement entries")
+    categorize_parser.add_argument('-d', required=False,
+                                   help="Delimiter to use when reading the"
+                                        " bank statement", default=',')
 
-    report_data = types.ReportData(category_overview=category_overview_df,
-                                   net_income=net_income_df,
-                                   total_value=total_value_df)
+    report_parser = subparsers.add_parser("report")
+    report_parser.set_defaults(func=report_func)
+    report_parser.add_argument('-i', required=True,
+                               help="CSV output file from categorization step")
 
-    document_builder.build_document(report_data)
+    args = parser.parse_args()
+
+    if hasattr(args, 'func'):
+        args.func(args)
+    else:
+        parser.print_help()
 
 
 if __name__ == "__main__":

banking_breakdown/document_builder.py

@@ -14,7 +14,11 @@ def _serialize_report_data(report_data: types.ReportData):
     report_data.net_income.to_csv('build/net_income.csv', index=False)
     report_data.category_overview.to_csv('build/category_overview.csv',
                                          index=False)
+    report_data.expenses_by_category.to_csv('build/expenses_by_category.csv',
+                                            index=False)
     report_data.total_value.to_csv('build/total_value.csv', index=False)
+    report_data.detailed_balance.to_csv('build/detailed_balance.csv',
+                                        index=False)
 
 
 def _compile_document():

banking_breakdown/regex_categorizer.py

@@ -0,0 +1,54 @@
+import pandas as pd
+import json
+
+
+def _is_str_column(s: pd.Series):
+    """Check if the type of a pandas DataFrame column is str.
+
+    Taken from https://stackoverflow.com/a/67001213/3433817.
+    """
+    if isinstance(s.dtype, pd.StringDtype):
+        # The series was explicitly created as a string series (Pandas>=1.0.0)
+        return True
+    elif s.dtype == 'object':
+        # Object series, check each value
+        return all((v is None) or isinstance(v, str) for v in s)
+    else:
+        return False
+
+
+def _read_regex_dict(regex_file: str):
+    with open(regex_file, 'r') as f:
+        return json.load(f)
+
+
+def assign_categories(df: pd.DataFrame, regex_file: str) -> pd.DataFrame:
+    if 'category' not in df.columns:
+        df['category'] = [' '] * len(df.index)
+
+    regex_dict = _read_regex_dict(regex_file)
+
+    df = df.fillna('')
+
+    for column in df.columns:
+        if not _is_str_column(df[column]):
+            continue
+
+        for category in regex_dict:
+            for regex in regex_dict[category]:
+                matched = df[column].str.contains(regex, regex=True)
+                df.loc[matched, 'category'] = category
+
+    return df
+
+
+def main():
+    df = pd.read_csv('../res/bank_statement_2023_categorized.csv')
+
+    df = assign_categories(df, regex_file='../res/regexes.json')
+
+    print(df['category'])
+
+
+if __name__ == "__main__":
+    main()

banking_breakdown/statement_parser.py

@@ -0,0 +1,122 @@
+import typing
+import pandas as pd
+from banking_breakdown import types
+import json
+import re
+import numpy as np
+
+
+def _escape_string(to_escape: str):
+    return to_escape.translate(str.maketrans({"&": r"\&"}))
+
+
+def _compute_total_balance(df: pd.DataFrame) -> pd.DataFrame:
+    stripped_df = pd.DataFrame(
+        {'t': df["t"], 'value': df["balance"]})
+
+    stripped_df.index = stripped_df['t']
+    gb = stripped_df.groupby(pd.Grouper(freq='M'))
+
+    result = gb.tail(1)['value'].reset_index()
+
+    return result
+
+
+def _compute_net_income(df: pd.DataFrame) -> pd.DataFrame:
+    df.index = df['t']
+    income_df = df.loc[df['value'] > 0]
+    expenses_df = df.loc[df['value'] < 0]
+
+    income_df = income_df.groupby(pd.Grouper(freq='M'))[
+        'value'].sum().reset_index().round(decimals=2)
+    expenses_df = expenses_df.groupby(pd.Grouper(freq='M'))[
+        'value'].sum().reset_index().round(decimals=2)
+
+    t = income_df['t']
+    income = income_df['value'].round(decimals=2)
+    expenses = expenses_df['value'].round(decimals=2)
+    net = (income + expenses).round(decimals=2)
+
+    result_df = pd.DataFrame(
+        {'t': t, 'income': income, 'expenses': expenses, 'net': net})
+
+    return result_df
+
+
+def _compute_category_overview(df: pd.DataFrame) -> pd.DataFrame:
+    df = df.loc[df['value'] < 0]
+    df = df.drop('t', axis=1)
+    df = df.groupby(['category']).sum().reset_index()
+    values = (df['value'] / df['value'].sum() * 100).to_numpy()
+    values[-1] += 100 - np.sum(values)
+    values = np.round(values, decimals=1)
+    categories = [_escape_string(category) for category in df['category']]
+
+    category_overview_df = pd.DataFrame(
+        {"category": categories, "value": values})
+
+    category_overview_df = category_overview_df.sort_values('value',
+                                                            ascending=False)
+
+    return category_overview_df
+
+
+def _compute_expenses_by_category(complete_df: pd.DataFrame) -> pd.DataFrame:
+    complete_df = complete_df.loc[complete_df['value'] < 0].copy()
+    complete_df['value'] = -complete_df['value']
+    complete_df.index = complete_df['t']
+
+    complete_gb = complete_df.groupby(pd.Grouper(freq='M'))
+
+    categories = complete_df['category'].unique()
+    data_dict = {category: [] for category in categories}
+
+    for (month_date, month_df) in complete_gb:
+        month_df = month_df.drop('t', axis=1).reset_index().drop('t', axis=1)
+
+        category_df = month_df.groupby(['category']).sum().reset_index()
+
+        for _, row in category_df.iterrows():
+            data_dict[row['category']].append(row['value'])
+
+        non_listed = list(set(categories) - set(category_df['category']))
+        for category in non_listed:
+            data_dict[category].append(0)
+
+    result = pd.DataFrame(data_dict)
+    result = result.reindex(result.mean().sort_values(ascending=False).index,
+                            axis=1)
+    result = result.round(decimals=2)
+    result['t'] = complete_gb.tail(1).drop('t', axis=1).reset_index()['t']
+
+    return result
+
+
+def _compute_detailed_balance(df: pd.DataFrame) -> pd.DataFrame:
+    return pd.DataFrame({'t': df["t"],
+                         'value': df["balance"]})
+
+
+def parse_statement(filename: str) -> types.ReportData:
+    df = pd.read_csv(filename)
+    df["t"] = pd.to_datetime(df["t"], format='%Y-%m-%d')
+
+    category_overview_df = _compute_category_overview(df)
+    total_balance_df = _compute_total_balance(df)
+    net_income_df = _compute_net_income(df)
+    detailed_balance_df = _compute_detailed_balance(df)
+    expenses_by_category_df = _compute_expenses_by_category(df)
+
+    return types.ReportData(category_overview_df,
+                            expenses_by_category_df,
+                            net_income_df,
+                            total_balance_df,
+                            detailed_balance_df, )
+
+
+def main():
+    report_data = parse_statement("../res/bank_statement_2023_categorized.csv")
+
+
+if __name__ == "__main__":
+    main()

banking_breakdown/test.json

@@ -0,0 +1,5 @@
+{
+  "asdf": [
+    "Kinemic"
+  ]
+}

banking_breakdown/types.py

@@ -5,5 +5,7 @@ import pandas as pd
 @dataclass
 class ReportData:
     category_overview: pd.DataFrame
+    expenses_by_category: pd.DataFrame
     net_income: pd.DataFrame
     total_value: pd.DataFrame
+    detailed_balance: pd.DataFrame

banking_breakdown/ui/__init__.py

@@ -0,0 +1,17 @@
+import sys
+
+import pandas as pd
+from PyQt6.QtWidgets import QApplication
+
+from banking_breakdown.ui.main_window import MainWindow
+
+
+def show_main_window(df: pd.DataFrame = None):
+    app = QApplication(sys.argv)
+    window = MainWindow()
+
+    if df is not None:
+        window.set_statement_data(df)
+
+    window.show()
+    app.exec()

banking_breakdown/ui/custom_ui_items.py

@@ -0,0 +1,161 @@
+import pandas as pd
+from PyQt6.QtCore import Qt
+from PyQt6.QtGui import QPixmap, QAction
+from PyQt6.QtWidgets import QHBoxLayout, QLabel, QMenu, QInputDialog, \
+    QMessageBox
+
+from banking_breakdown.ui.pandas_model import PandasModel
+
+
+class WarningItem(QHBoxLayout):
+    """Item appearing at top of Window with warning icon."""
+
+    def __init__(self, text: str, parent=None):
+        super(WarningItem, self).__init__()
+
+        self._warningIcon = QLabel()
+        pixmap = QPixmap("res/warning.png")
+        self._warningIcon.setPixmap(pixmap)
+
+        self._label = QLabel(text)
+        self._label.setWordWrap(True)
+
+        self.addWidget(self._warningIcon)
+        self.addWidget(self._label)
+
+        self.setStretch(0, 0)
+        self.setStretch(1, 1)
+
+    def hide(self):
+        self._label.hide()
+        self._warningIcon.hide()
+
+
+class HeaderContextMenu(QMenu):
+    """Context menu appearing when right-clicking the header of the QTableView.
+    """
+
+    def __init__(self, column_index, pandas_model: PandasModel, callback=None,
+                 parent=None):
+        super(HeaderContextMenu, self).__init__()
+
+        self._pandas_model = pandas_model
+        self._callback = callback
+
+        self._column_index = column_index
+        self._column_text \
+            = self._pandas_model.headerData(self._column_index,
+                                            Qt.Orientation.Horizontal)
+
+        # Define assign action
+
+        assign_menu = QMenu("Assign type", self)
+        assign_date_action = QAction("date", self)
+        assign_float_action = QAction("float", self)
+
+        assign_menu.addAction(assign_date_action)
+        assign_menu.addAction(assign_float_action)
+
+        assign_date_action.triggered.connect(self._assign_date_handler)
+        assign_float_action.triggered.connect(self._assign_float_handler)
+
+        # Define other actions
+
+        rename_action = QAction("Rename", self)
+        delete_action = QAction("Delete", self)
+        switch_action = QAction("Switch position", self)
+
+        rename_action.triggered.connect(self._rename_handler)
+        delete_action.triggered.connect(self._delete_handler)
+        switch_action.triggered.connect(self._switch_handler)
+
+        # Add actions to menu
+
+        self.addAction(rename_action)
+        self.addAction(delete_action)
+        self.addAction(switch_action)
+        self.addAction(assign_menu.menuAction())
+
+    def _rename_handler(self):
+        new_name, flag = QInputDialog.getText(self, "Rename column",
+                                              "New name:",
+                                              text=self._column_text)
+
+        if not flag:
+            return
+
+        if (new_name != self._column_text) and (new_name != ''):
+            try:
+                self._pandas_model.rename_column(self._column_text, new_name)
+            except:
+                QMessageBox.warning(self, "No action performed",
+                                    "An error occurred.")
+
+            if self._callback:
+                self._callback()
+
+    def _delete_handler(self):
+        button = QMessageBox.question(self, "Delete column",
+                                      f"Are you sure you want to delete"
+                                      f" column '{self._column_text}'?")
+
+        if button == QMessageBox.StandardButton.Yes:
+            self._pandas_model.delete_column_by_index(self._column_index)
+
+            if self._callback:
+                self._callback()
+
+    def _switch_handler(self):
+        df = self._pandas_model.get_dataframe()
+        columns = [column for column in df.columns
+                   if column != self._column_text]
+
+        other_name, flag = QInputDialog.getItem(self, "Switch column position",
+                                                f"Switch position of colum "
+                                                f"'{self._column_text}' with:",
+                                                columns, editable=False)
+
+        if not flag:
+            return
+
+        self._pandas_model.switch_columns(self._column_text, other_name)
+
+        if self._callback:
+            self._callback()
+
+    def _assign_date_handler(self):
+        date_format, flag = QInputDialog.getText(self, "Format",
+                                                 "Format:",
+                                                 text="%d.%m.%Y")
+
+        if not flag:
+            return
+
+        try:
+            self._pandas_model.assign_date_column(self._column_text,
+                                                  date_format)
+        except:
+            QMessageBox.warning(self, "No action performed",
+                                "An error occurred.")
+
+        if self._callback:
+            self._callback()
+
+    def _assign_float_handler(self):
+        chars = ['.', ',']
+        decimal_sep, flag = QInputDialog.getItem(self, "Decimal separator",
+                                                 "Decimal separator:",
+                                                 chars, editable=False)
+
+        if not flag:
+            return
+
+        try:
+            self._pandas_model.assign_float_column(self._column_text,
+                                                   decimal_sep)
+        except:
+            QMessageBox.warning(self, "No action performed",
+                                "An error occurred.")
+
+        if self._callback:
+            self._callback()

banking_breakdown/ui/main_window.py

@@ -0,0 +1,245 @@
+import typing
+import pandas as pd
+from PyQt6 import uic
+from PyQt6.QtCore import Qt, QSortFilterProxyModel
+from PyQt6.QtGui import QAction
+from PyQt6.QtWidgets import QMainWindow, QPushButton, QVBoxLayout, \
+    QTableView, QInputDialog, QMessageBox, QFileDialog, QListWidget
+
+from banking_breakdown.ui.pandas_model import PandasModel
+from banking_breakdown.ui.custom_ui_items import WarningItem, HeaderContextMenu
+
+
+class MainWindow(QMainWindow):
+    def __init__(self):
+        super(MainWindow, self).__init__()
+
+        uic.loadUi("res/main_window.ui", self)
+        self.setContentsMargins(9, 9, 9, 9)
+
+        self._warnings = []
+
+        # Extract elements
+
+        self._warning_layout \
+            = self.findChild(QVBoxLayout, "warningLayout")
+        self._create_button \
+            = self.findChild(QPushButton, "createCategoryButton")
+        self._delete_button \
+            = self.findChild(QPushButton, "deleteCategoryButton")
+        self._apply_button \
+            = self.findChild(QPushButton, "applyCategoryButton")
+        self._clear_button \
+            = self.findChild(QPushButton, "clearCategoryButton")
+        self._list_widget \
+            = self.findChild(QListWidget, "categoryListWidget")
+        self._table_view \
+            = self.findChild(QTableView, "transactionTableView")
+        self._action_save \
+            = self.findChild(QAction, "actionSave")
+
+        # Set up QTableView model
+
+        self._pandas_model = PandasModel(self)
+        self._proxy_model = QSortFilterProxyModel(self)
+        self._proxy_model.setSourceModel(self._pandas_model)
+        self._table_view.setModel(self._proxy_model)
+        self._proxy_model.setSortRole(Qt.ItemDataRole.EditRole)
+        self._proxy_model.setDynamicSortFilter(False)
+
+        # Set event handlers
+
+        self._create_button.clicked.connect(self._handle_create_click)
+        self._delete_button.clicked.connect(self._handle_delete_click)
+        self._clear_button.clicked.connect(self._handle_clear_click)
+        self._apply_button.clicked.connect(self._handle_apply_click)
+        self._list_widget.itemActivated.connect(self._handle_item_double_click)
+
+        self._action_save.triggered.connect(self._handle_save)
+
+        header = self._table_view.horizontalHeader()
+        header.setContextMenuPolicy(Qt.ContextMenuPolicy.CustomContextMenu)
+        header.customContextMenuRequested.connect(
+            self._handle_header_right_click)
+
+        self._list_widget.itemSelectionChanged.connect(
+            self._handle_list_selection_changed)
+
+        self._table_view.selectionModel().selectionChanged.connect(
+            self._handle_table_selection_changed)
+
+    #
+    # Table data updates
+    #
+
+    def set_statement_data(self, df: pd.DataFrame):
+        self._pandas_model.set_dataframe(df)
+        self._dataframe_update_callback()
+        self._resize_table_columns_to_content()
+
+    def get_statement_data(self) -> pd.DataFrame:
+        return self._pandas_model.get_dataframe()
+
+    def _dataframe_update_callback(self):
+        self._show_warnings()
+        self._update_categories_from_dataframe()
+
+    def _resize_table_columns_to_content(self):
+        df = self._pandas_model.get_dataframe()
+
+        # Resize columns
+
+        if len(df.columns) < 10:  # Experimentally determined threshold
+            # Properly resize columns (takes longer)
+            self._table_view.resizeColumnsToContents()
+        else:
+            # Quickly approximate sizes
+            for i, col in enumerate(df.columns):
+                max_char = max(max([len(str(x)) for x in df[col].values]),
+                               len(col))
+                self._table_view.setColumnWidth(i, max_char * 10)
+
+    def _assign_category_to_selected_transactions(self, category: str):
+        indexes = self._table_view.selectionModel().selectedRows()
+        row_indices = [self._table_view.model().mapToSource(index).row()
+                       for index in indexes]
+
+        self._pandas_model.assign_category(category, row_indices)
+
+    #
+    # List data updates
+    #
+
+    def _add_categories(self, categories: typing.Sequence[str]):
+        for category in categories:
+            self._list_widget.addItem(category)
+
+    def _update_categories_from_dataframe(self):
+        df_categories = self._pandas_model.get_categories()
+        current_categories = [self._list_widget.item(x).text() for x
+                              in range(self._list_widget.count())]
+        missing = list(set(df_categories) - set(current_categories))
+
+        self._add_categories([category for category
+                              in missing if category != ' '])
+
+    #
+    # Warnings
+    #
+
+    def _add_warning_item(self, text: str):
+        warning_item = WarningItem(text=text, parent=self)
+
+        self._warning_layout.addLayout(warning_item)
+        self._warnings.append(warning_item)
+
+    def _show_warnings(self):
+        for warning_item in self._warnings:
+            warning_item.hide()
+            self._warning_layout.removeItem(warning_item)
+
+        columns = self._pandas_model.get_columns()
+
+        if 't' not in columns:
+            self._add_warning_item(
+                "The column 't' does not exist. Please rename the column"
+                " containing the dates of the transactions to 't'.")
+
+        if 'value' not in columns:
+            self._add_warning_item(
+                "The column 'value' does not exist. Please rename the column"
+                " containing the values of the transactions to 'value'.")
+
+        if 'balance' not in columns:
+            self._add_warning_item(
+                "The column 'balance' does not exist. Please rename the column"
+                " containing the balance after each transaction to 'balance'")
+
+    #
+    # Event handlers
+    #
+
+    def _handle_header_right_click(self, pos):
+        column = self._table_view.horizontalHeader().logicalIndexAt(pos)
+
+        context = HeaderContextMenu(parent=self, column_index=column,
+                                    pandas_model=self._pandas_model,
+                                    callback=self._dataframe_update_callback)
+        context.exec(self.sender().mapToGlobal(pos))
+
+    def _handle_create_click(self):
+        new_name, flag = QInputDialog.getText(self, "Create category",
+                                              "New category:",
+                                              text="Category")
+        if not flag:
+            return
+
+        current_items = [self._list_widget.item(x).text() for x
+                         in range(self._list_widget.count())]
+
+        if new_name not in current_items:
+            self._add_categories([new_name])
+        else:
+            QMessageBox.warning(self, "No action performed",
+                                f"Category '{new_name}' already exists.")
+
+    def _handle_delete_click(self):
+        selected_item = self._list_widget.selectedItems()[0]
+
+        button = QMessageBox.question(self, "Delete category",
+                                      f"Are you sure you want to delete"
+                                      f" category '{selected_item.text()}'?")
+
+        if button == QMessageBox.StandardButton.Yes:
+            self._pandas_model.delete_category(selected_item.text())
+            self._list_widget.takeItem(self._list_widget.row(selected_item))
+
+    def _handle_clear_click(self):
+        self._assign_category_to_selected_transactions(' ')
+
+    def _handle_apply_click(self):
+        category = self._list_widget.selectedItems()[0].text()
+        self._assign_category_to_selected_transactions(category)
+
+    def _handle_item_double_click(self, item):
+        self._assign_category_to_selected_transactions(item.text())
+
+    def _handle_save(self):
+        filename, _ = QFileDialog.getSaveFileName(self, 'Save File')
+
+        if filename == '':
+            return
+
+        df = self.get_statement_data()
+        df.to_csv(filename, index=False)
+
+    #
+    # Enable / Disable buttons
+    #
+
+    def _check_enable_delete_button(self):
+        if len(self._list_widget.selectedItems()) > 0:
+            self._delete_button.setEnabled(True)
+        else:
+            self._delete_button.setEnabled(False)
+
+    def _check_enable_clear_button(self):
+        if len(self._table_view.selectionModel().selectedRows()) > 0:
+            self._clear_button.setEnabled(True)
+        else:
+            self._clear_button.setEnabled(False)
+
+    def _check_enable_apply_button(self):
+        if ((len(self._table_view.selectionModel().selectedRows()) > 0)
+                and (len(self._list_widget.selectedItems()) > 0)):
+            self._apply_button.setEnabled(True)
+        else:
+            self._apply_button.setEnabled(False)
+
+    def _handle_list_selection_changed(self):
+        self._check_enable_delete_button()
+        self._check_enable_apply_button()
+
+    def _handle_table_selection_changed(self):
+        self._check_enable_clear_button()
+        self._check_enable_apply_button()

banking_breakdown/ui/pandas_model.py

@@ -0,0 +1,172 @@
+import typing
+
+import numpy
+import pandas as pd
+from PyQt6 import QtCore
+from PyQt6.QtCore import Qt, QModelIndex, QSortFilterProxyModel
+
+
+def _get_str_dataframe(df: pd.DataFrame) -> pd.DataFrame:
+    """Return a given dataframe with all values turned into strings.
+
+    When the data given to the PandasModel class contains non-strings,
+    an attached QTableView seems to respond rather slowly. This function
+    turns all data in the DataFrame into strings, yielding a better experience.
+    """
+    return df.astype(str)
+
+
+class PandasModel(QtCore.QAbstractTableModel):
+    def __init__(self, parent=None):
+        QtCore.QAbstractTableModel.__init__(self, parent)
+
+        self._data = pd.DataFrame()
+        self._data_str = pd.DataFrame()
+
+    #
+    # Overloaded functions
+    #
+
+    def rowCount(self, parent=None):
+        return len(self._data_str.values)
+
+    def columnCount(self, parent=None):
+        return self._data_str.columns.size
+
+    def data(self, index, role=Qt.ItemDataRole.DisplayRole):
+        if not index.isValid():
+            return QtCore.QVariant()
+
+        if (role != Qt.ItemDataRole.DisplayRole) and (
+                role != Qt.ItemDataRole.EditRole):
+            return QtCore.QVariant()
+
+        if role == Qt.ItemDataRole.DisplayRole:
+            item = self._data_str.iloc[index.row(), index.column()]
+            return QtCore.QVariant(item)
+        elif role == Qt.ItemDataRole.EditRole:
+            item = self._data.iloc[index.row(), index.column()]
+            if type(item) is numpy.float64:
+                return QtCore.QVariant(float(item))
+            else:
+                return QtCore.QVariant(item)
+
+    def headerData(self, section, orientation,
+                   role=Qt.ItemDataRole.DisplayRole):
+
+        if not ((orientation == Qt.Orientation.Horizontal)
+                and (role == Qt.ItemDataRole.DisplayRole)):
+            return super().headerData(section, orientation, role)
+
+        return self._data_str.columns[section]
+
+    #
+    # Manipulate categories
+    #
+
+    def assign_category(self, category, row_indices):
+        if 'category' not in self._data.columns:
+            self.create_column('category')
+
+        self._data.loc[row_indices, 'category'] = category
+        self._data_str = _get_str_dataframe(self._data)
+
+        for row_index in row_indices:
+            start_index = self.index(row_index, 0)
+            stop_index = self.index(row_index, len(self._data.columns) - 1)
+            self.dataChanged.emit(start_index, stop_index)
+
+    def delete_category(self, category):
+        if 'category' not in self._data.columns:
+            self.create_column('category')
+
+        row_indices = self._data.loc[self._data['category'] == category].index
+        self.assign_category(' ', row_indices)
+
+    def get_categories(self) -> typing.List[str]:
+        if 'category' not in self._data.columns:
+            self.create_column('category')
+
+        return self._data['category'].unique()
+
+    #
+    # Manipulate columns
+    #
+
+    def create_column(self, column, initial_value=' '):
+        self._data[column] = [initial_value] * len(self._data.index)
+        self._data_str = _get_str_dataframe(self._data)
+        self.layoutAboutToBeChanged.emit()
+        self.layoutChanged.emit()
+
+    def delete_column_by_index(self, column_index):
+        self._data \
+            = self._data.iloc[:, [j for j, c in enumerate(self._data.columns)
+                                  if j != column_index]]
+        self._data_str = _get_str_dataframe(self._data)
+
+        self.layoutAboutToBeChanged.emit()
+        self.layoutChanged.emit()
+
+    def rename_column(self, old_name, new_name):
+        if new_name in self._data.columns:
+            raise Exception(
+                f"A column with the name '{new_name}' already exists.")
+
+        self._data = self._data.rename(columns={old_name: new_name})
+        self._data_str = _get_str_dataframe(self._data)
+
+        column_index = self._data.columns.get_loc(new_name)
+        self.headerDataChanged.emit(Qt.Orientation.Horizontal,
+                                    column_index, column_index)
+
+    def switch_columns(self, column1, column2):
+        column_titles = list(self._data.columns)
+
+        index1, index2 \
+            = column_titles.index(column1), column_titles.index(column2)
+
+        column_titles[index1], column_titles[index2] \
+            = column_titles[index2], column_titles[index1]
+
+        self._data = self._data.reindex(columns=column_titles)
+        self._data_str = _get_str_dataframe(self._data)
+
+        self.layoutAboutToBeChanged.emit()
+        self.layoutChanged.emit()
+
+    def get_columns(self) -> typing.List[str]:
+        return list(self._data.columns)
+
+    def assign_float_column(self, column, decimal_sep):
+        if decimal_sep == ',':
+            self._data[column] \
+                = self._data[column].str.replace(',', '.').astype(float)
+        else:
+            self._data[column] = self._data[column].astype(float)
+
+        self._data_str = _get_str_dataframe(self._data)
+
+        column_index = self._data.columns.get_loc(column)
+        start_index = self.index(0, column_index)
+        stop_index = self.index(len(self._data.index), column_index)
+        self.dataChanged.emit(start_index, stop_index)
+
+    def assign_date_column(self, column, date_format):
+        self._data[column] \
+            = pd.to_datetime(self._data[column], format=date_format)
+        self._data_str = _get_str_dataframe(self._data)
+
+    #
+    # Directly access dataframe
+    #
+
+    def set_dataframe(self, df):
+        self._data = df
+        self._data_str = _get_str_dataframe(df)
+
+        self.layoutAboutToBeChanged.emit()
+        self.layoutChanged.emit()
+
+    def get_dataframe(self) -> pd.DataFrame:
+        return self._data

res/main_window.ui

@@ -0,0 +1,176 @@
+<?xml version="1.0" encoding="UTF-8"?>
+<ui version="4.0">
+ <class>MainWindow</class>
+ <widget class="QMainWindow" name="MainWindow">
+  <property name="geometry">
+   <rect>
+    <x>0</x>
+    <y>0</y>
+    <width>1300</width>
+    <height>731</height>
+   </rect>
+  </property>
+  <property name="windowTitle">
+   <string>MainWindow</string>
+  </property>
+  <widget class="QWidget" name="centralwidget">
+   <layout class="QVBoxLayout" name="verticalLayout_4">
+    <property name="spacing">
+     <number>0</number>
+    </property>
+    <item>
+     <layout class="QHBoxLayout" name="horizontalLayout_6">
+      <item>
+       <widget class="QWidget" name="widget" native="true">
+        <property name="maximumSize">
+         <size>
+          <width>600</width>
+          <height>16777215</height>
+         </size>
+        </property>
+        <layout class="QVBoxLayout" name="warningLayout">
+         <property name="topMargin">
+          <number>0</number>
+         </property>
+         <property name="bottomMargin">
+          <number>0</number>
+         </property>
+        </layout>
+       </widget>
+      </item>
+     </layout>
+    </item>
+    <item>
+     <layout class="QHBoxLayout" name="horizontalLayout_3" stretch="0,1">
+      <item>
+       <widget class="QGroupBox" name="groupBox">
+        <property name="title">
+         <string>Categories</string>
+        </property>
+        <layout class="QVBoxLayout" name="verticalLayout_2">
+         <item>
+          <layout class="QHBoxLayout" name="horizontalLayout">
+           <item>
+            <widget class="QPushButton" name="createCategoryButton">
+             <property name="toolTip">
+              <string>Create new category</string>
+             </property>
+             <property name="text">
+              <string>Create</string>
+             </property>
+            </widget>
+           </item>
+           <item>
+            <widget class="QPushButton" name="deleteCategoryButton">
+             <property name="enabled">
+              <bool>false</bool>
+             </property>
+             <property name="toolTip">
+              <string>Delete selected category</string>
+             </property>
+             <property name="text">
+              <string>Delete</string>
+             </property>
+            </widget>
+           </item>
+          </layout>
+         </item>
+         <item>
+          <widget class="QListWidget" name="categoryListWidget">
+           <property name="sizeAdjustPolicy">
+            <enum>QAbstractScrollArea::AdjustToContents</enum>
+           </property>
+          </widget>
+         </item>
+         <item>
+          <layout class="QHBoxLayout" name="horizontalLayout_2">
+           <item>
+            <widget class="QPushButton" name="clearCategoryButton">
+             <property name="enabled">
+              <bool>false</bool>
+             </property>
+             <property name="toolTip">
+              <string>Clear category field of selected transactions</string>
+             </property>
+             <property name="text">
+              <string>Clear</string>
+             </property>
+            </widget>
+           </item>
+           <item>
+            <widget class="QPushButton" name="applyCategoryButton">
+             <property name="enabled">
+              <bool>false</bool>
+             </property>
+             <property name="toolTip">
+              <string>Apply selected category to selected transactions</string>
+             </property>
+             <property name="text">
+              <string>Apply</string>
+             </property>
+            </widget>
+           </item>
+          </layout>
+         </item>
+        </layout>
+       </widget>
+      </item>
+      <item>
+       <widget class="QGroupBox" name="groupBox_2">
+        <property name="title">
+         <string>Transactions</string>
+        </property>
+        <layout class="QVBoxLayout" name="verticalLayout">
+         <item>
+          <widget class="QTableView" name="transactionTableView">
+           <property name="selectionBehavior">
+            <enum>QAbstractItemView::SelectRows</enum>
+           </property>
+           <property name="sortingEnabled">
+            <bool>true</bool>
+           </property>
+           <attribute name="horizontalHeaderShowSortIndicator" stdset="0">
+            <bool>true</bool>
+           </attribute>
+           <attribute name="horizontalHeaderStretchLastSection">
+            <bool>true</bool>
+           </attribute>
+          </widget>
+         </item>
+        </layout>
+       </widget>
+      </item>
+     </layout>
+    </item>
+   </layout>
+  </widget>
+  <widget class="QMenuBar" name="menubar">
+   <property name="geometry">
+    <rect>
+     <x>0</x>
+     <y>0</y>
+     <width>1300</width>
+     <height>23</height>
+    </rect>
+   </property>
+   <widget class="QMenu" name="menuFile">
+    <property name="title">
+     <string>File</string>
+    </property>
+    <addaction name="actionSave"/>
+   </widget>
+   <addaction name="menuFile"/>
+  </widget>
+  <widget class="QStatusBar" name="statusbar"/>
+  <action name="actionSave">
+   <property name="text">
+    <string>Save</string>
+   </property>
+   <property name="shortcut">
+    <string>Ctrl+S</string>
+   </property>
+  </action>
+ </widget>
+ <resources/>
+ <connections/>
+</ui>

res/report.tex

@@ -4,6 +4,7 @@
 \usepackage{amsmath}
 \usepackage{pgfplots}
 \pgfplotsset{compat=newest}
+\usetikzlibrary{pgfplots.dateplot}
 
 % Other packages
 \usepackage[a4paper, total={12cm, 25cm}]{geometry}
@@ -30,6 +31,10 @@
 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
 
 
+\makeatletter
+\newcommand*\shortyear[1]{\expandafter\@gobbletwo\number\numexpr#1\relax}
+\makeatother
+
 \newcommand{\slice}[6]{
 	\pgfmathparse{0.5*#1+0.5*#2}
 	\let\midangle\pgfmathresult
@@ -136,16 +141,41 @@
 
 		\begin{tikzpicture}
 			\begin{axis}[
+				date coordinates in=x,
 				width=\textwidth,
 				height=0.375\textwidth,
 				ylabel={Net income in €},
 				y label style={at={(-0.1,0.5)},anchor=south},
+				xticklabel=\month.\shortyear{\year},
+				xtick=data,
+				xticklabel style={
+					rotate=60,
+					anchor=near xticklabel,
+				},
+				grid,
+				enlarge x limits=0.03,
 			]
-				\addplot+[ybar, color=scol2, fill=scol2, line width=1pt]
-					table[col sep=comma, x=t, y=value, discard if lt={value}{0}]
+				% Dummy plot to set x axis ticks
+				\addplot[draw=none]
+					table[col sep=comma, x=t, y=net]
 						{net_income.csv};
-				\addplot+[ybar, color=scol0, fill=scol0, line width=1pt]
-					table[col sep=comma, x=t, y=value, discard if gt={value}{0}]
+
+				% Dummy plot to set x axis scale
+				\addplot[draw=none]
+					table[col sep=comma, x=t, y expr=0]
+						{detailed_balance.csv};
+
+				\addplot[ybar, bar width=0.4cm, draw=none, fill=scol2!30, line width=1pt]
+					table[col sep=comma, x=t, y=income]
+						{net_income.csv};
+				\addplot[ybar, bar width=0.4cm, draw=none, fill=scol0!30, line width=1pt]
+					table[col sep=comma, x=t, y=expenses]
+						{net_income.csv};
+                \addplot[ybar, bar width=0.3cm, draw=none, fill=scol0, line width=1pt]
+					table[col sep=comma, x=t, y=net, discard if gt={net}{0}]
+						{net_income.csv};
+                \addplot[ybar, bar width=0.3cm, draw=none, fill=scol2, line width=1pt]
+					table[col sep=comma, x=t, y=net, discard if lt={net}{0}]
 						{net_income.csv};
 			\end{axis}
 		\end{tikzpicture}
@@ -155,15 +185,31 @@
 
 		\begin{tikzpicture}
 			\begin{axis}[
+				date coordinates in=x,
 				width=\textwidth,
 				height=0.375\textwidth,
-				area style,
 				ylabel={Total balance in €},
 				y label style={at={(-0.1,0.5)},anchor=south},
+				xticklabel=\month.\shortyear{\year},
+				xtick=data,
+				enlarge x limits=0.03,
+				xticklabel style={
+					rotate=60,
+					anchor=near xticklabel,
+				},
+				grid,
 			]
-				\addplot+[mark=none, color=scol1, line width=1pt]
+				% Dummy plot to set x axis ticks
+				\addplot[draw=none]
 					table[col sep=comma, x=t, y=value]
-						{total_value.csv} \closedcycle;
+						{total_value.csv};
+
+				\addplot[scol3, mark=none, line width=1pt]
+					table[col sep=comma, x=t, y=value]
+						{detailed_balance.csv};
+				\addplot[scol1, mark=none, line width=1pt]
+					table[col sep=comma, x=t, y=value]
+						{total_value.csv};
 			\end{axis}
 		\end{tikzpicture}
 	\end{subfigure}
@@ -171,5 +217,105 @@
 	\caption{Development of account balance over time}
 \end{figure}
 
+\begin{figure}
+	\centering
+
+	
+	\csvautotabular{net_income.csv}
+\end{figure}
+
+
+\begin{figure}[H]
+	\centering
+
+	% Read table
+	\pgfplotstableread[col sep=comma]{expenses_by_category.csv}\expbycattable
+	\pgfplotstablegetcolsof{\expbycattable}
+	\pgfmathtruncatemacro\NumCols{\pgfplotsretval-1}
+
+	\begin{subfigure}[c]{\textwidth}
+		\centering
+
+		\begin{tikzpicture}
+			\begin{axis}[
+				stack plots=y,
+				area style,
+				date coordinates in=x,
+				width=\textwidth,
+				height=0.375\textwidth,
+				xticklabel=\month.\shortyear{\year},
+				xtick=data,
+				enlargelimits=false,
+				xticklabel style={
+					rotate=60,
+					anchor=near xticklabel,
+				},
+				legend columns=5,
+				legend style={at={(0.5,-0.6)},anchor=south},
+				ylabel={Expenses in €},
+				ymin=0,
+			]
+				% For each
+				\pgfplotsinvokeforeach{0,...,\NumCols/2 -1}{
+					% Define color
+					\pgfmathparse{1000 / (\NumCols/2 -1) * #1}
+					\extractcolormapcolor{tempcol#1}{\pgfmathresult}
+
+					% Add plot
+					\addplot+[tempcol#1]
+						table[col sep=comma, x=t, y index=#1]
+							{\expbycattable} \closedcycle;
+
+					% Add legend entry (https://tex.stackexchange.com/a/405018)
+					\pgfplotstablegetcolumnnamebyindex{#1}\of{\expbycattable}\to\pgfplotsretval
+					\expandafter\addlegendentry\expandafter{\pgfplotsretval}
+				}
+			\end{axis}
+		\end{tikzpicture}
+	\end{subfigure}\\[1em]
+	\begin{subfigure}[c]{\textwidth}
+		\centering
+
+		\begin{tikzpicture}
+			\begin{axis}[
+				stack plots=y,
+				area style,
+				date coordinates in=x,
+				width=\textwidth,
+				height=0.375\textwidth,
+				xticklabel=\month.\shortyear{\year},
+				xtick=data,
+				enlargelimits=false,
+				xticklabel style={
+					rotate=60,
+					anchor=near xticklabel,
+				},
+				legend columns=5,
+				legend style={at={(0.5,-0.6)},anchor=south},
+				ylabel={Expenses in €},
+				ymin=0,
+			]
+				% For each
+				\pgfplotsinvokeforeach{\NumCols/2,...,\NumCols-1}{
+					% Define color
+					\pgfmathparse{1000 * (#1 - \NumCols/2) / (\NumCols-1 - \NumCols/2)}
+					\extractcolormapcolor{tempcol#1}{\pgfmathresult}
+
+					% Add plot
+					\addplot+[tempcol#1]
+						table[col sep=comma, x=t, y index=#1]
+							{\expbycattable} \closedcycle;
+
+					% Add legend entry (https://tex.stackexchange.com/a/405018)
+					\pgfplotstablegetcolumnnamebyindex{#1}\of{\expbycattable}\to\pgfplotsretval
+					\expandafter\addlegendentry\expandafter{\pgfplotsretval}
+				}
+			\end{axis}
+		\end{tikzpicture}
+	\end{subfigure}
+
+	\caption{Expenses by category}
+\end{figure}
+
 \end{document}