Turn a messy sales table into three charts that make an argument

The situation

A clothing brand in Karachi has twelve months of sales across four cities and five product lines. The founder shows investors a dashboard of fifteen default charts and they leave confused. She needs three charts that each make one clear point, and she wants to know you can find the story before you draw it.

Explore the data to find the argument, then build three honest, readable charts.

What you'll be able to show

• Write the claim a chart makes before drawing it
• Choose the right library and chart type for each point
• Strip a chart to what carries information
• Use charts to explore, then deliver only the few that argue

What you need

• Python with pandas, matplotlib and seaborn, in any notebook
• A sales dataset you build: month, city, product_line, revenue, units, for twelve months

Tasks

1. 1Build the data

   Create at least 240 rows (12 months x 4 cities x 5 lines) with revenue and units. Put a real story in it on purpose: one city declining, one product line rising sharply, one month that is an outlier.
   A correct result: The dataset and a note of the three patterns you planted.

2. 2Explore first

   Plot the distribution of revenue and units, and a simple trend per city. These are for you, not the report. Write down what you actually found, including whether your planted story is the clearest one.
   A correct result: The exploratory charts and a short paragraph of what the data shows.

3. 3Write the three claims

   Decide the three sentences your report will make, each backed by one chart. Each must be a claim, not a topic ('Lahore revenue fell every quarter', not 'Lahore sales').
   A correct result: Three claim sentences.

4. 4Draw them honestly

   Build one chart per claim, choosing the type and library deliberately. Order bars by value, start axes at zero where a bar chart requires it, and label directly rather than relying on a legend.
   A correct result: Three charts, each matching its claim, with a note on why you chose each type.

5. 5Strip each one

   Remove everything that carries no information: unused gridlines, borders, duplicate legends, excess decimals. Put the claim sentence as the title of each chart.
   A correct result: The before-and-after of one chart and the final three with sentence titles.

6. 6Deliver as an argument

   Arrange the three in an order that builds, with the finding written beside each. Show it to someone without explaining, and note every point where they hesitated.
   A correct result: The three-chart report and a list of the hesitations you observed.

What to hand in

The notebook with the exploration, the three claims, the final three charts with sentence titles in a built order, and the notes from watching someone read it.

How it is marked