|
| 1 | +"""Interactive Streamlit data-cleaning recipe. |
| 2 | +
|
| 3 | +Install dependencies with:: |
| 4 | +
|
| 5 | + pip install streamlit freshdata-cleaner |
| 6 | +
|
| 7 | +Run the app with:: |
| 8 | +
|
| 9 | + streamlit run examples/integrations/streamlit_app.py |
| 10 | +""" |
| 11 | + |
| 12 | +from __future__ import annotations |
| 13 | + |
| 14 | +from io import BytesIO |
| 15 | +from typing import Any |
| 16 | + |
| 17 | +import pandas as pd |
| 18 | +import streamlit as st |
| 19 | + |
| 20 | +import freshdata as fd |
| 21 | + |
| 22 | + |
| 23 | +def demo_data() -> pd.DataFrame: |
| 24 | + """Return a small messy dataset for trying the app without an upload.""" |
| 25 | + return pd.DataFrame( |
| 26 | + { |
| 27 | + "Customer Name": [" Ava Chen ", "Noah Patel", "Mia Lopez", "Noah Patel"], |
| 28 | + "Orders": ["3", "N/A", "5", "N/A"], |
| 29 | + "Region": ["West", "East", "West", "East"], |
| 30 | + } |
| 31 | + ) |
| 32 | + |
| 33 | + |
| 34 | +def load_uploaded_csv(contents: bytes) -> pd.DataFrame: |
| 35 | + """Read uploaded CSV bytes into a DataFrame.""" |
| 36 | + return pd.read_csv(BytesIO(contents)) |
| 37 | + |
| 38 | + |
| 39 | +def count_modified_columns(report: Any) -> int: |
| 40 | + """Count columns modified by applied cleaning actions, imputations, or drops.""" |
| 41 | + modified = set(report.columns_imputed) | set(report.columns_dropped) |
| 42 | + modified.update( |
| 43 | + action.column |
| 44 | + for action in report.actions |
| 45 | + if action.column is not None |
| 46 | + and action.count > 0 |
| 47 | + and getattr(action, "status", "automatic") in {"automatic", "approved"} |
| 48 | + ) |
| 49 | + return len(modified) |
| 50 | + |
| 51 | + |
| 52 | +def main() -> None: |
| 53 | + st.set_page_config(page_title="FreshData Cleaning Lab", layout="wide") |
| 54 | + st.title("FreshData Cleaning Lab") |
| 55 | + st.caption( |
| 56 | + "Upload a CSV to compare its original data with FreshData's cleaned output." |
| 57 | + ) |
| 58 | + |
| 59 | + uploaded_file = st.file_uploader("Choose a CSV file", type=["csv"]) |
| 60 | + |
| 61 | + if uploaded_file is None: |
| 62 | + raw_df = demo_data() |
| 63 | + st.info("Showing sample data. Upload a CSV to clean your own dataset.") |
| 64 | + else: |
| 65 | + try: |
| 66 | + raw_df = load_uploaded_csv(uploaded_file.getvalue()) |
| 67 | + except (ValueError, UnicodeDecodeError, pd.errors.ParserError) as exc: |
| 68 | + st.error(f"Could not read this CSV: {exc}") |
| 69 | + st.stop() |
| 70 | + |
| 71 | + st.sidebar.header("Cleaning Settings") |
| 72 | + strategy = st.sidebar.selectbox( |
| 73 | + "Strategy", ["balanced", "conservative", "aggressive"], index=0 |
| 74 | + ) |
| 75 | + drop_duplicates = st.sidebar.checkbox("Drop duplicate rows", value=True) |
| 76 | + |
| 77 | + try: |
| 78 | + cleaned_df, report = fd.clean( |
| 79 | + raw_df, |
| 80 | + strategy=strategy, |
| 81 | + drop_duplicates=drop_duplicates, |
| 82 | + return_report=True, |
| 83 | + ) |
| 84 | + except Exception as exc: |
| 85 | + st.error(f"FreshData could not clean this dataset: {exc}") |
| 86 | + st.stop() |
| 87 | + |
| 88 | + st.subheader("Cleaning summary") |
| 89 | + metric_columns = st.columns(4) |
| 90 | + metric_columns[0].metric( |
| 91 | + "Rows removed", max(report.rows_before - report.rows_after, 0) |
| 92 | + ) |
| 93 | + metric_columns[1].metric("Columns modified", count_modified_columns(report)) |
| 94 | + imputed_values = sum( |
| 95 | + action.count |
| 96 | + for action in report.actions |
| 97 | + if ( |
| 98 | + action.step == "impute" |
| 99 | + or ( |
| 100 | + action.step == "missing" |
| 101 | + and action.column in report.columns_imputed |
| 102 | + ) |
| 103 | + ) |
| 104 | + and getattr(action, "status", "automatic") in {"automatic", "approved"} |
| 105 | + and action.count > 0 |
| 106 | + ) |
| 107 | + metric_columns[2].metric("Missing values imputed", imputed_values) |
| 108 | + metric_columns[3].metric("Duplicates dropped", report.duplicates_removed) |
| 109 | + |
| 110 | + raw_column, cleaned_column = st.columns(2) |
| 111 | + with raw_column: |
| 112 | + st.subheader("Raw data") |
| 113 | + st.dataframe(raw_df, use_container_width=True) |
| 114 | + with cleaned_column: |
| 115 | + st.subheader("Cleaned data") |
| 116 | + st.dataframe(cleaned_df, use_container_width=True) |
| 117 | + |
| 118 | + st.download_button( |
| 119 | + "Download cleaned CSV", |
| 120 | + data=cleaned_df.to_csv(index=False).encode("utf-8"), |
| 121 | + file_name="cleaned_data.csv", |
| 122 | + mime="text/csv", |
| 123 | + ) |
| 124 | + |
| 125 | + with st.expander("Quality report"): |
| 126 | + st.code(report.summary(), language="text") |
| 127 | + if report.actions: |
| 128 | + st.dataframe( |
| 129 | + pd.DataFrame( |
| 130 | + [ |
| 131 | + { |
| 132 | + "Step": action.step, |
| 133 | + "Column": action.column or "", |
| 134 | + "Change": action.description, |
| 135 | + "Count": action.count, |
| 136 | + } |
| 137 | + for action in report.actions |
| 138 | + ] |
| 139 | + ), |
| 140 | + use_container_width=True, |
| 141 | + ) |
| 142 | + |
| 143 | + |
| 144 | +if __name__ == "__main__": |
| 145 | + main() |
0 commit comments