Function squarify

missingno/missingno.py:731–748

(_min_x, _max_x, _min_y, _max_y, df)

        # Recursive quadtree. This subroutine, when, builds a dictionary of squares, stored by tuples keyed with
        # (min_x, max_x, min_y, max_y), whose values are the nullity of squares containing less than 100 observations.
        def squarify(_min_x, _max_x, _min_y, _max_y, df):
            arr = df[[x_col, y_col]].values
            points_inside = df[(_min_x < arr[:,0]) &
                               (arr[:,0] < _max_x) &
                               (_min_y < arr[:,1]) &
                               (arr[:,1] < _max_y)]
            if len(points_inside) < cutoff:
                # The following subroutine groups `geo_group` by `x_col` and `y_col`, and calculates and returns
                # a list of points in the group (`points`) as well as its overall nullity (`geographic_nullity`). The
                # first of these calculations is ignored.
                _, square_nullity = _calculate_geographic_nullity(points_inside, x_col, y_col)
                rectangles.append(((_min_x, _max_x,_min_y, _max_y), square_nullity))
            else:
                _mid_x, _mid_y = (_min_x + _max_x) / 2, (_min_y + _max_y) / 2
                squarify(_min_x, _mid_x, _mid_y, _max_y, points_inside)
                squarify(_min_x, _mid_x, _min_y, _mid_y, points_inside)
                squarify(_mid_x, _max_x, _mid_y, _max_y, points_inside)
                squarify(_mid_x, _max_x, _min_y, _mid_y, points_inside)

        # Populate the `squares` array, per the above.
        squarify(min_x, max_x, min_y, max_y, df)