diff --git a/README.md b/README.md
index eedd9cb4..dab96da2 100644
--- a/README.md
+++ b/README.md
@@ -71,6 +71,9 @@ Alphalens depends on:
 - [seaborn](https://github.com/mwaskom/seaborn)
 - [statsmodels](https://github.com/statsmodels/statsmodels)
 
+> Note that Numpy>=2.0 requires pandas>=2.2.2. If you are using an older version of pandas, you may need to upgrade
+> accordingly, otherwise you may encounter compatibility issues.
+
 # Usage
 
 A good way to get started is to run the examples in a [Jupyter notebook](https://jupyter.org/).
diff --git a/pyproject.toml b/pyproject.toml
index 429bf94b..28d4cd4c 100644
--- a/pyproject.toml
+++ b/pyproject.toml
@@ -30,9 +30,11 @@ classifiers = [
 ]
 
 dependencies = [
+    # following pandas
+    "numpy>=1.23.5; python_version<'3.12'",
+    "numpy>=1.26.0; python_version>='3.12'",
+    "pandas >=1.3.0,<3.0",
     "matplotlib >=1.4.0",
-    "numpy >=1.9.1",
-    "pandas >=1.2.0,<3.0",
     "scipy >=0.14.0",
     "seaborn >=0.6.0",
     "statsmodels >=0.6.1",
@@ -49,8 +51,6 @@ documentation = 'https://alphalens.ml4trading.io'
 requires = [
     'setuptools>=54.0.0',
     "setuptools_scm[toml]>=6.2",
-    'wheel>=0.31.0',
-    'oldest-supported-numpy; python_version>="3.8"',
 ]
 
 build-backend = 'setuptools.build_meta'
@@ -98,12 +98,12 @@ version_scheme = 'guess-next-dev'
 local_scheme = 'dirty-tag'
 
 
-[tool.pytest]
+[tool.pytest.ini_options]
+pythonpath = ['src']
 minversion = "6.0"
 testpaths = 'tests'
 addopts = '-v'
 
-
 [tool.cibuildwheel]
 test-extras = "test"
 test-command = "pytest -n 2 {package}/tests"
@@ -122,7 +122,7 @@ skip = "*musllinux*"
 
 [tool.black]
 line-length = 88
-target-version = ['py38', 'py39', 'py310']
+target-version = ['py39', 'py310', 'py311', 'py312']
 include = '\.pyi?$'
 extend-exclude = '''
 \(
@@ -130,10 +130,21 @@ extend-exclude = '''
 \)
 '''
 
+
 [tool.tox]
 legacy_tox_ini = """
 [tox]
-envlist = py39-pandas12, py{39,310,311}-pandas{13,14,15}, py{39,310,311,312}-pandas{15,20,21,22}
+
+envlist =
+    py39-pandas{13,14,15}-numpy1
+    py310-pandas{13,14,15,20,21,22}-numpy1
+    py311-pandas{13,14,15,20,21,22}-numpy1
+    py312-pandas{13,14,15,20,21,22}-numpy1
+    py39-pandas222-numpy2
+    py310-pandas222-numpy2
+    py311-pandas222-numpy2
+    py312-pandas222-numpy2
+
 isolated_build = True
 skip_missing_interpreters = True
 minversion = 3.23.0
@@ -153,13 +164,16 @@ setenv =
 changedir = tmp
 extras = test
 deps =
-    pandas12: pandas>=1.2.0,<1.3
     pandas13: pandas>=1.3.0,<1.4
     pandas14: pandas>=1.4.0,<1.5
     pandas15: pandas>=1.5.0,<1.6
     pandas20: pandas>=2.0,<2.1
     pandas21: pandas>=2.1,<2.2
     pandas22: pandas>=2.2,<2.3
+    pandas222: pandas>=2.2.2,<2.3
+    numpy1: numpy>=1.23.5,<2.0
+    numpy2: numpy>=2.0,<2.1
+
 
 commands =
     pytest -n 2 --cov={toxinidir}/src --cov-report term  --cov-report=xml --cov-report=html:htmlcov {toxinidir}/tests
diff --git a/src/alphalens/performance.py b/src/alphalens/performance.py
index 665fe6a7..0e467086 100644
--- a/src/alphalens/performance.py
+++ b/src/alphalens/performance.py
@@ -70,7 +70,7 @@ def src_ic(group):
     if by_group:
         grouper.append("group")
 
-    ic = factor_data.groupby(grouper).apply(src_ic)
+    ic = factor_data.groupby(grouper, observed=True).apply(src_ic)
     if by_group:
         return ic
     else:
@@ -195,9 +195,9 @@ def to_weights(group, _demeaned, _equal_weight):
     if group_adjust:
         grouper.append("group")
 
-    weights = factor_data.groupby(grouper, group_keys=False)["factor"].apply(
-        to_weights, demeaned, equal_weight
-    )
+    weights = factor_data.groupby(grouper, group_keys=False, observed=True)[
+        "factor"
+    ].apply(to_weights, demeaned, equal_weight)
 
     if group_adjust:
         weights = weights.groupby(level="date", group_keys=False).apply(
@@ -507,7 +507,7 @@ def mean_return_by_quantile(
     if by_group:
         grouper.append("group")
 
-    group_stats = factor_data.groupby(grouper)[
+    group_stats = factor_data.groupby(grouper, observed=True)[
         utils.get_forward_returns_columns(factor_data.columns)
     ].agg(["mean", "std", "count"])
 
@@ -847,7 +847,7 @@ def average_cumulative_return(q_fact, demean_by):
         #
         returns_bygroup = []
 
-        for group, g_data in factor_data.groupby("group"):
+        for group, g_data in factor_data.groupby("group", observed=True):
             g_fq = g_data["factor_quantile"]
             if group_adjust:
                 demean_by = g_fq  # demeans at group level
@@ -878,7 +878,7 @@ def average_cumulative_return(q_fact, demean_by):
         #
         if group_adjust:
             all_returns = []
-            for group, g_data in factor_data.groupby("group"):
+            for group, g_data in factor_data.groupby("group", observed=True):
                 g_fq = g_data["factor_quantile"]
                 avgcumret = g_fq.groupby(g_fq).apply(
                     cumulative_return_around_event, g_fq