Skip to content

Streamlit Lab 2 — CSV Data Viewer with Filters

Week: Module C (Streamlit module) · CLO-1, CLO-3 · Time: ~75 min

Learning objectives

  • Read an uploaded CSV with pd.read_csv(uploaded).
  • Display overview metrics with st.metric and data with st.dataframe.
  • Filter by columns (multiselect) and values (selectbox) interactively.
  • Handle the "no file uploaded yet" first run.

Problem statement

Build one app (app.py) — a CSV viewer that:

  1. Asks the user to upload a CSV.
  2. Shows overview metrics (rows, columns, missing values).
  3. Shows the data with a column filter (multiselect of columns).
  4. Lets the user pick one numeric column and see its describe() summary and mean as a metric.

Test it with the tips dataset: in a notebook run import seaborn as sns; sns.load_dataset("tips").to_csv("tips.csv") (or use any CSV from the course's datasets/).

Tasks

  1. Create streamlit-lab2/app.py.
  2. Upload + guard: 
    uploaded = st.file_uploader("Upload a CSV", type=["csv"])
if uploaded is None:
    st.info("Upload a CSV to begin.")
    st.stop()
df = pd.read_csv(uploaded)
  3. Overview: three st.metric cards — rows, columns, missing values.
  4. Column filter: 
    cols = st.multiselect("Show columns", df.columns, default=list(df.columns))
if cols:
    st.dataframe(df[cols])
else:
    st.info("Pick at least one column.")
  5. Numeric summary: a selectbox over numeric columns + st.metric with the mean + st.dataframe(df[col].describe()).
  6. Run the app, upload tips.csv, verify each interaction, screenshot.

Expected output (verify all)

  • Before upload: upload box + info message, nothing crashes.
  • After uploading tips.csv: Rows = 244, Columns = 7, Missing = 0.
  • Unchecking a column removes it from the table; empty selection shows the info message instead of an error.
  • Picking tip as the numeric column shows mean ≈ 2.998 and the full describe() table.

Questions to answer in your submission:

  1. Why must the app check uploaded is None before reading the file?
  2. What does st.multiselect return when the user removes all columns, and why does if cols: prevent a crash?

Challenge

Add a search box (st.text_input) that filters the dataframe by substring across all columns (hint: df.astype(str).apply(lambda row: row.str.contains(q, case=False).any(), axis=1)), and show the filtered row count as a metric so the user sees how many rows match.