diff --git a/Dockerfile b/Dockerfile
new file mode 100644
index 0000000..211a8c0
--- /dev/null
+++ b/Dockerfile
@@ -0,0 +1,11 @@
+FROM python:3.12.2-slim-bullseye
+
+WORKDIR /app
+
+COPY app /app
+
+RUN pip3 install -r requirements.txt
+
+EXPOSE 8501
+
+ENTRYPOINT ["streamlit", "run", "streamlit_app.py", "--server.port=8501", "--server.address=0.0.0.0"]
\ No newline at end of file
diff --git a/app/requirements.txt b/app/requirements.txt
new file mode 100644
index 0000000..5fa91e6
--- /dev/null
+++ b/app/requirements.txt
@@ -0,0 +1,5 @@
+pandas==2.2.0
+streamlit==1.31.0
+plotly==5.19.0
+sqlalchemy==2.0.27
+psycopg2-binary==2.9.9
diff --git a/app/streamlit_app.py b/app/streamlit_app.py
new file mode 100644
index 0000000..89dd36c
--- /dev/null
+++ b/app/streamlit_app.py
@@ -0,0 +1,301 @@
+import os
+import pandas as pd
+import streamlit as st
+import plotly.express as px
+from sqlalchemy import create_engine
+from sqlalchemy.engine import URL
+
+
+st.set_page_config(
+    page_title="GêXis Gaúcho",
+    page_icon="🍔",
+    layout="wide",
+)
+
+
+@st.cache_data
+def load_data():
+    url_object = URL.create(
+        "postgresql+psycopg2",
+        username=os.environ["LOGIN"],
+        password=os.environ["PASSWORD"],
+        host=os.environ["HOST"],
+        port=os.environ["PORT"],
+        database=os.environ["DATABASE"],
+    )
+    engine = create_engine(url_object)
+    df = pd.read_sql_table(
+        "ge_validations_store_normalized", con=engine, schema="great_expectations"
+    )
+    df["meta.validation_time"] = pd.to_datetime(
+        df["meta.validation_time"], format="%Y%m%dT%H%M%S.%fZ"
+    )
+    df["meta.validation_date"] = df["meta.validation_time"].apply(lambda x: x.date())
+    df["meta.validation_date"] = pd.to_datetime(df["meta.validation_date"])
+    df["meta.validation_yearmonth"] = df["meta.validation_time"].apply(
+        lambda x: pd.Timestamp(f"{x.year}-{x.month}-01")
+    )
+    df["schema_table_name"] = (
+        df["meta.batch_spec.schema_name"] + "." + df["meta.batch_spec.table_name"]
+    )
+
+    return df
+
+
+data = load_data()
+
+"""
+# 🍔 GêXis Gaúcho
+---
+"""
+
+col1, col2 = st.columns([1.5, 2.5])
+
+with col1:
+    """
+    ## Últimas validation :red[failed]
+    """
+
+    last_run = data.groupby(["schema_table_name"])["meta.validation_time"].max()
+    last_run = last_run.reset_index()
+    last_run_full = last_run.merge(
+        data,
+        on=[
+            "schema_table_name",
+            "meta.validation_time",
+        ],
+        how="inner",
+    )
+
+    st.dataframe(
+        last_run_full[~last_run_full["success"]].sort_values(
+            by="meta.validation_time", ascending=False
+        ),
+        column_order=[
+            "meta.validation_time",
+            "schema_table_name",
+        ],
+        column_config={
+            "schema_table_name": st.column_config.TextColumn(
+                "schema.table",
+            ),
+            "meta.validation_time": st.column_config.DatetimeColumn(
+                "Validation date",
+                format="DD/MM/YYYY",
+            ),
+        },
+        hide_index=True,
+    )
+
+
+with col2:
+    """
+    ## Percentual de sucesso por dia
+    """
+
+    count_success = (
+        data.groupby("meta.validation_date")["success"]
+        .value_counts()
+        .unstack(fill_value=0)
+    )
+    count_success["success_percentual"] = count_success[True] / (
+        count_success[True] + count_success[False]
+    )
+    count_success = count_success.reset_index()
+
+    fig1 = px.bar(count_success, x="meta.validation_date", y="success_percentual")
+    fig1.update_layout(
+        width=800,
+    )
+    fig1.update_traces(marker_color="#F63366", marker_line_width=0)
+
+    st.plotly_chart(fig1, theme="streamlit")
+
+"""
+---
+"""
+
+col3, col4 = st.columns([1.5, 2.5])
+
+with col3:
+    """
+    ## Top-10 `schema.table` errors
+    """
+
+    top_fail = (
+        data[~data["success"]]
+        .groupby(["schema_table_name"])["success"]
+        .value_counts()
+        .unstack(fill_value=0)
+        .sort_values(False, ascending=False)
+        .head(10)
+    )
+    top_fail = top_fail.reset_index()
+
+    st.dataframe(
+        top_fail,
+        column_config={
+            "schema_table_name": st.column_config.TextColumn(
+                "schema.table",
+            ),
+            "False": st.column_config.NumberColumn(
+                "Qt. Erros",
+                format="%d",
+            ),
+        },
+    )
+
+with col4:
+    """
+    ## Top-10 `schema.table` errors por mês
+    """
+
+    top_fail_time = (
+        data[~data["success"]]
+        .groupby(["schema_table_name", "meta.validation_yearmonth"])["success"]
+        .count()
+    )
+    top_fail_time = top_fail_time.reset_index()
+
+    top_fail_time_pivot = top_fail_time[
+        top_fail_time["schema_table_name"].isin(top_fail["schema_table_name"])
+    ].pivot_table(
+        index="meta.validation_yearmonth",
+        columns="schema_table_name",
+        values="success",
+        fill_value=0,
+    )
+    top_fail_time_pivot = top_fail_time_pivot.reset_index()
+
+    fig2 = px.line(
+        top_fail_time_pivot,
+        x="meta.validation_yearmonth",
+        y=top_fail_time_pivot.columns,
+        #XXX check this
+        hover_data={"meta.validation_yearmonth": "|%B %d, %Y"},
+    )
+    #XXX check this
+    fig2.update_xaxes(dtick="M1", tickformat="%b\n%Y")
+    fig2.update_layout(
+        width=800,
+    )
+
+    st.plotly_chart(fig2, theme="streamlit")
+
+"""
+---
+
+## Raw data
+"""
+
+col5, col6, col7, col8 = st.columns(4)
+
+with col5:
+    which_schema = st.selectbox(
+        "Schema",
+        options=sorted(data["meta.batch_spec.schema_name"].unique()),
+        index=None,
+    )
+
+with col6:
+    which_table = st.selectbox(
+        "Table",
+        options=(
+            sorted(
+                data.loc[
+                    data["meta.batch_spec.schema_name"] == which_schema,
+                    "meta.batch_spec.table_name",
+                ].unique()
+            )
+            if which_schema
+            else sorted(data["meta.batch_spec.table_name"].unique())
+        ),
+        index=None,
+    )
+
+with col7:
+    which_date_min, which_date_max = st.date_input(
+        "Validation date range",
+        (data["meta.validation_time"].min(), data["meta.validation_time"].max()),
+        format="DD/MM/YYYY",
+    )
+
+with col8:
+    is_success = st.multiselect("is_success", [True, False], [True, False])
+
+st.markdown(
+    f"""
+    You selected: `{which_schema}`, `{which_table}`, `{which_date_min}`, `{which_date_max}`, `{is_success}`
+"""
+)
+
+if which_schema and not which_table:
+    df_raw = data[
+        (data["meta.batch_spec.schema_name"] == which_schema)
+        & (data["meta.validation_time"].dt.date >= which_date_min)
+        & (data["meta.validation_time"].dt.date <= which_date_max)
+        & (data["success"].isin(is_success))
+    ]
+elif which_table:
+    df_raw = data[
+        (data["meta.batch_spec.table_name"] == which_table)
+        & (data["meta.validation_time"].dt.date >= which_date_min)
+        & (data["meta.validation_time"].dt.date <= which_date_max)
+        & (data["success"].isin(is_success))
+    ]
+else:
+    df_raw = data[
+        (data["meta.validation_time"].dt.date >= which_date_min)
+        & (data["meta.validation_time"].dt.date <= which_date_max)
+        & (data["success"].isin(is_success))
+    ]
+
+st.dataframe(
+    df_raw,
+    column_order=[
+        "success",
+        "meta.validation_time",
+        "meta.batch_spec.schema_name",
+        "meta.batch_spec.table_name",
+        "meta.active_batch_definition.datasource_name",
+        "expectation_config.expectation_type",
+        "expectation_config.kwargs.max_value",
+        "expectation_config.kwargs.min_value",
+        "result.observed_value",
+    ],
+    column_config={
+        "success": st.column_config.TextColumn(
+            "Rodou?",
+        ),
+        "meta.validation_time": st.column_config.DatetimeColumn(
+            "Validation time",
+            format="DD/MM/YYYY",
+        ),
+        "meta.batch_spec.schema_name": st.column_config.TextColumn(
+            "Schema",
+        ),
+        "meta.batch_spec.table_name": st.column_config.TextColumn(
+            "Table",
+        ),
+        "meta.active_batch_definition.datasource_name": st.column_config.TextColumn(
+            "Conn",
+        ),
+        "expectation_config.expectation_type": st.column_config.TextColumn(
+            "Expectation type",
+        ),
+        "expectation_config.kwargs.max_value": st.column_config.NumberColumn(
+            "Max value allowed",
+            format="%d",
+        ),
+        "expectation_config.kwargs.min_value": st.column_config.NumberColumn(
+            "Min value allowed",
+            format="%d",
+        ),
+        "result.observed_value": st.column_config.NumberColumn(
+            "Value observed",
+            format="%d",
+        ),
+    },
+    width=1400,
+    hide_index=True,
+)
diff --git a/docker-compose.yml b/docker-compose.yml
new file mode 100644
index 0000000..3a9ac0c
--- /dev/null
+++ b/docker-compose.yml
@@ -0,0 +1,22 @@
+version: '3'
+
+services:
+  streamlit:
+    build:
+      context: .
+    environment:
+      LOGIN: ""
+      PASSWORD: ""
+      HOST: ""
+      PORT: ""
+      DATABASE: ""
+    ports:
+      - 8501:8501
+    volumes:
+      - ./app:/app
+    healthcheck:
+      test: ["CMD", "curl", "--fail", "http://localhost:8501/_stcore/health"]
+      interval: 10s
+      timeout: 10s
+      retries: 5
+    restart: always