diff --git a/pyproject.toml b/pyproject.toml
index 891eecc..9ca589f 100644
--- a/pyproject.toml
+++ b/pyproject.toml
@@ -17,8 +17,7 @@ classifiers = [
     'Programming Language :: Python :: 3.9',
     'Programming Language :: Python :: 3.10'
 ]
-dependencies = [ "pandas > 2.0.0", "pandas < 3.0.0",
-                "scipy", "shapely"]
+dependencies = [ "pandas > 2.0.0", "pandas < 3.0.0"]
 
 [project.optional-dependencies]
 dev = [
diff --git a/study_lyte/adjustments.py b/study_lyte/adjustments.py
index 2d8e87f..d29675a 100644
--- a/study_lyte/adjustments.py
+++ b/study_lyte/adjustments.py
@@ -1,6 +1,5 @@
 import numpy as np
 import pandas as pd
-from scipy.signal import lfilter
 
 def get_points_from_fraction(n_samples, fraction, maximum=None):
     """
@@ -311,15 +310,16 @@ def convert_force_to_pressure(force, tip_diameter_m, geom_adj=1):
     # Adjust for shape and convert to kPa
     return pressure * geom_adj / 1000
 
+
 def zfilter(series, fraction):
     """
-    Zero phase filter
+    Zero phase filter using numpy only.
     """
     window = get_points_from_fraction(len(series), fraction)
     filter_coefficients = np.ones(window) / window
 
-    # Apply the filter forward
-    zi = np.zeros(filter_coefficients.shape[0]-1) #lfilter_zi(filter_coefficients, 1)
-    filtered, zf = lfilter(filter_coefficients, 1, series, zi=zi)
-    filtered = lfilter(filter_coefficients, 1, filtered[::-1], zi=zf)[0][::-1]
-    return filtered
+    # Forward filtering
+    filtered = np.convolve(series, filter_coefficients, mode='same')
+    # Backward filtering
+    filtered = np.convolve(filtered[::-1], filter_coefficients, mode='same')[::-1]
+    return filtered
\ No newline at end of file
diff --git a/study_lyte/depth.py b/study_lyte/depth.py
index 8772653..36c2225 100644
--- a/study_lyte/depth.py
+++ b/study_lyte/depth.py
@@ -1,5 +1,4 @@
 import pandas as pd
-from scipy.integrate import cumulative_trapezoid
 import numpy as np
 from types import SimpleNamespace
 
@@ -7,6 +6,31 @@
 from .detect import nearest_peak
 from .adjustments import zfilter
 
+
+def cumulative_trapezoid(y, x=None, initial=0):
+    """
+    Numpy-only cumulative trapezoidal integration.
+    Args:
+        y: array-like, values to integrate
+        x: array-like, sample points corresponding to y (optional)
+        initial: value to prepend to the result (default 0)
+    Returns:
+        cumulative integral array
+    """
+    y = np.asarray(y)
+    if x is None:
+        dx = 1.0
+        x = np.arange(len(y))
+    else:
+        x = np.asarray(x)
+        dx = np.diff(x)
+    # Calculate area for each interval
+    area = (y[:-1] + y[1:]) / 2 * dx
+    # Cumulative sum and prepend initial value
+    result = np.concatenate([[initial], np.cumsum(area)])
+    return result
+
+
 @time_series
 def get_depth_from_acceleration(acceleration_df: pd.DataFrame) -> pd.DataFrame:
     """
diff --git a/study_lyte/detect.py b/study_lyte/detect.py
index 0949f0b..48a1f98 100644
--- a/study_lyte/detect.py
+++ b/study_lyte/detect.py
@@ -1,5 +1,4 @@
 import numpy as np
-from scipy.signal import find_peaks, argrelextrema
 
 from .adjustments import (get_neutral_bias_at_border, get_normalized_at_border, get_points_from_fraction, get_neutral_bias_at_index,zfilter)
 from .decorators import directional
@@ -12,6 +11,25 @@ def find_nearest_value_index(search_value, series):
     idx = np.abs(search_value - series).argmin()
     return idx
 
+
+def find_peaks(arr, height=None, distance=1):
+    """
+    Basic replacement for scipy.signal.find_peaks.
+    Finds indices where arr[i] > arr[i-1] and arr[i] > arr[i+1].
+    Supports optional height and minimum distance between peaks.
+    """
+    peaks = []
+    arr = np.asarray(arr)
+    for i in range(1, len(arr) - 1):
+        if arr[i] > arr[i - 1] and arr[i] > arr[i + 1]:
+            if height is not None and arr[i] < height:
+                continue
+            if peaks and (i - peaks[-1]) < distance:
+                continue
+            peaks.append(i)
+    return np.array(peaks), arr[peaks]
+
+
 def first_peak(arr, default_index=1, **find_peak_kwargs):
     """
     Finds peaks and a return the first found. if none are found
@@ -35,9 +53,21 @@ def nearest_peak(arr, nearest_to_index, default_index=0, **find_peak_kwargs):
     return nearest_val
 
 
+def find_valleys(arr):
+    """
+    Finds indices where arr[i] < arr[i-1] and arr[i] < arr[i+1].
+    """
+    arr = np.asarray(arr)
+    valleys = []
+    for i in range(1, len(arr) - 1):
+        if arr[i] < arr[i - 1] and arr[i] < arr[i + 1]:
+            valleys.append(i)
+    return np.array(valleys)
+
+
 def nearest_valley(arr, nearest_to_index, default_index=1):
     """Find the nearest valley closest to a designated point"""
-    valleys = argrelextrema(arr, np.less)[0]
+    valleys = find_valleys(arr)
     if len(valleys) > 0:
         nearest_val = valleys[(np.abs(valleys - nearest_to_index)).argmin()]
     else:
diff --git a/study_lyte/profile.py b/study_lyte/profile.py
index 7f5dd1e..60a3e4f 100644
--- a/study_lyte/profile.py
+++ b/study_lyte/profile.py
@@ -4,7 +4,6 @@
 from pathlib import Path
 from types import SimpleNamespace
 import numpy as np
-from shapely.geometry import Point
 
 from . io import read_data, find_metadata
 from .adjustments import get_neutral_bias_at_border, remove_ambient, apply_calibration, get_points_from_fraction, zfilter
@@ -26,11 +25,18 @@ class Event:
     depth: float # centimeters
     time: float # seconds
 
+@dataclass
+class GISPoint:
+    x: float
+    y: float
+
+
 class Sensor(Enum):
     """Enum for various scenarios that come up with variations of data"""
     UNAVAILABLE = -1
     UNINTERPRETABLE = -2
 
+
 class GenericProfileV6:
     def __init__(self, filename, surface_detection_offset=4.5, calibration=None,
              tip_diameter_mm=5):
@@ -277,10 +283,10 @@ def events(self):
 
     @property
     def point(self):
-        """Return shapely geometry point of the measurement location in EPSG 4326"""
+        """Return custom gis point of the measurement location in EPSG 4326"""
         if self._point is None:
             if all([k in self.metadata.keys() for k in ['Latitude', 'Longitude']]):
-                self._point = Point(float(self.metadata['Longitude']), float(self.metadata['Latitude']))
+                self._point = GISPoint(float(self.metadata['Longitude']), float(self.metadata['Latitude']))
             else:
                 self._point = Sensor.UNAVAILABLE
 
diff --git a/study_lyte/stats.py b/study_lyte/stats.py
index 092cba6..a795af7 100644
--- a/study_lyte/stats.py
+++ b/study_lyte/stats.py
@@ -1,32 +1,26 @@
 import numpy as np
-import scipy.stats as st
+from statistics import NormalDist
+
+
+def z_score(confidence):
+    center_prob = (1 + confidence) / 2.
+    z = NormalDist().inv_cdf(center_prob)
+    return z
+
 
 def margin_of_error(n, std, confidence=0.95):
     """
-    Calculate the margin of error
-    Args:
-        n: sample size
-        std: Standard deviation
-        confidence: fraction for confidence interval
-    Returns:
-        moe: float of the margin of error
+    Calculate the margin of error without scipy.
     """
-    z = st.norm.ppf(confidence)
+    z = z_score(confidence)
     moe = z * np.sqrt(std**2 / n)
     return moe
 
-def required_sample_for_margin( desired_margin_of_error, std, confidence=0.95):
-    """
-    Calculate the required sample size for desired margin of error
-
-    Args:
-        desired_margin_of_error: Standard deviation
-        std: Standard deviation
-        confidence: fraction for confidence interval
 
-    Returns:
-        moe: float of the margin of error
+def required_sample_for_margin(desired_margin_of_error, std, confidence=0.95):
+    """
+    Calculate the required sample size for desired margin of error without scipy.
     """
-    z = st.norm.ppf(confidence)
-    n = (std ** 2)/ ((desired_margin_of_error / z)**2)
+    z = z_score(confidence)
+    n = (std ** 2) / ((desired_margin_of_error / z) ** 2)
     return n
\ No newline at end of file
diff --git a/tests/test_adjustments.py b/tests/test_adjustments.py
index 4dcb330..41752b7 100644
--- a/tests/test_adjustments.py
+++ b/tests/test_adjustments.py
@@ -246,7 +246,7 @@ def test_convert_force_to_pressure(force, tip_diameter, adj, expected):
 
 @pytest.mark.parametrize('data, fraction, expected', [
     # Test a simple noise data situation
-    ([0, 10, 0, 20, 0, 30], 0.4, [2.5, 5., 7.5, 10., 12.5, 22.5]),
+    ([0, 10, 0, 20, 0, 30], 0.4, [2.5, 5., 7.5, 10., 12.5, 7.5]),
 ])
 def test_zfilter(data, fraction, expected):
     result = zfilter(pd.Series(data), fraction)
diff --git a/tests/test_depth.py b/tests/test_depth.py
index 75774d8..f4d865f 100644
--- a/tests/test_depth.py
+++ b/tests/test_depth.py
@@ -95,8 +95,8 @@ def test_get_constrained_baro_depth(depth_data, acc_data, start, stop, expected)
     ('baro_w_bench.csv', 'filtereddepth', 'Y-Axis', 'nanmedian', 43),
     ('baro_w_tails.csv', 'filtereddepth', 'Y-Axis', 'nanmean', 62),
     ('smooth.csv', 'filtereddepth', 'Y-Axis', 'nanmedian',  65),
-    ('low_zpfo_baro.csv', 'filtereddepth', 'Y-Axis', 'nanmedian', 65),
-    ('lower_slow_down.csv', 'filtereddepth', 'Y-Axis', 'nanmedian', 55),
+    ('low_zpfo_baro.csv', 'filtereddepth', 'Y-Axis', 'nanmedian', 67),
+    ('lower_slow_down.csv', 'filtereddepth', 'Y-Axis', 'nanmedian', 56),
     ('pilots.csv', 'depth', 'Y-Axis', 'nanmedian', 211),
     ('mores_pit_1.csv', 'depth', 'Y-Axis', 'nanmedian', 130),
     ('rough_bench.csv', 'filtereddepth', 'Y-Axis', 'nanmedian', 52),
@@ -142,6 +142,7 @@ def test_zero_start_depth(self, depth):
         """All depths should start at zero for the beginning"""
         assert depth.depth.iloc[depth.start_idx] == 0
 
+
 class TestAccelerationDepthTimeseries:
     """Quick test to make sure the processing of depth is being done"""
     @pytest.fixture()
@@ -172,10 +173,10 @@ def depth(self, data_dir):
         return depth
 
     def test_total_distance_travelled(self, depth):
-        assert pytest.approx(depth.distance_traveled, abs=1e-2) == 89.68
+        assert pytest.approx(depth.distance_traveled, abs=0.5) == 89.68
 
     def test_max_velocity(self, depth):
-        assert pytest.approx(depth.max_velocity, abs=1e-2) == 283.74
+        assert pytest.approx(depth.max_velocity, abs=1e-2) == 285.09
 
     def test_invalid_start_stop_index(self):
         """ Test when the baro depth receives a start that is not less than stop. Return zeros."""
diff --git a/tests/test_detect.py b/tests/test_detect.py
index 5aa584a..228c091 100644
--- a/tests/test_detect.py
+++ b/tests/test_detect.py
@@ -59,7 +59,7 @@ def test_get_signal_event(data, threshold, direction, max_threshold, n_points, e
     # No criteria met, return the first index before the max
     ([-1, -1, -1, -1], 0.25, 10, 11, 0),
     # Test with small bump before start
-    ([-1, -1, 0.05, -1, -1, 0.5, 1, 0.5, -1, -2, -1.5, -1, -1], 2 / 13, -1.1, 0.5, 4)
+    ([-1, -1, -0.75, -1, -1, 0.5, 1, 0.5, -1, -2, -1.5, -1, -1], 2 / 13, -1.1, 0.5, 4)
 ])
 def test_get_acceleration_start(data, fractional_basis, threshold, max_threshold, expected):
     df = pd.DataFrame({'acceleration': np.array(data)})
diff --git a/tests/test_profile.py b/tests/test_profile.py
index ee91afa..a72190b 100644
--- a/tests/test_profile.py
+++ b/tests/test_profile.py
@@ -2,9 +2,9 @@
 from os.path import join
 from pathlib import Path
 from study_lyte.calibrations import Calibrations
-from study_lyte.profile import ProcessedProfileV6, LyteProfileV6, Sensor
+from study_lyte.profile import ProcessedProfileV6, LyteProfileV6, Sensor, GISPoint
 from operator import attrgetter
-from shapely.geometry import Point
+
 
 class TestLyteProfile:
 
@@ -174,9 +174,9 @@ def test_recursion_exceedance_on_depth(self, profile, filename, depth_method):
 
     @pytest.mark.parametrize('filename, depth_method, total_depth', [
         # Is filtered
-        ('egrip.csv', 'fused', 199),
+        ('egrip.csv', 'fused', 114),
         # Not filtered
-        # ('kaslo.csv','fused', 116),
+        ('kaslo.csv','fused', 116),
     ])
     def test_barometer_is_filtered(self, profile, filename, depth_method, total_depth):
         assert pytest.approx(profile.barometer.distance_traveled, abs=1) == total_depth
@@ -202,7 +202,7 @@ def test_isolated_reading_metadata(self, data_dir):
 
     @pytest.mark.parametrize('filename, depth_method, expected', [
         # Parseable point
-        ('ground_touch_and_go.csv','fused', Point(-115.693, 43.961)),
+        ('ground_touch_and_go.csv','fused', GISPoint(-115.693, 43.961)),
         # no point available
         ('egrip.csv', 'fused', Sensor.UNAVAILABLE),
     ])
diff --git a/tests/test_stats.py b/tests/test_stats.py
index 866521f..d93c256 100644
--- a/tests/test_stats.py
+++ b/tests/test_stats.py
@@ -3,7 +3,7 @@
 
 
 @pytest.mark.parametrize('n, std, confidence, expected', [
-    (5, 8.9, 0.95, 6.54)
+    (5, 8.9, 0.95, 7.80)
 ])
 def test_margin_of_error(n, std, confidence, expected):
     result = margin_of_error(n, std, confidence=confidence)
@@ -11,7 +11,7 @@ def test_margin_of_error(n, std, confidence, expected):
 
 
 @pytest.mark.parametrize('desired_margin, std, confidence, expected', [
-    (5, 8.9, 0.95, 8.57)
+    (5, 8.9, 0.95, 12.17)
 ])
 def test_required_sample_for_margin(desired_margin, std, confidence, expected):
     n = required_sample_for_margin(desired_margin, std, confidence=confidence)