Business Intelligence & Reporting

1KPIs, metrics & framing the question

A KPI (Key Performance Indicator) is a metric tied to a goal. Good BI starts not with charts but with the question the business needs answered.

A dashboard usually leads with a few headline KPIs, then a trend, then a breakdown:

2Power BI: data modelling & DAX

Power BI is Microsoft's BI tool. Two skills make you productive: modelling your tables into a star schema, and writing DAX measures.

Total Sales = SUM ( Sales[Amount] )

Sales LY = CALCULATE ( [Total Sales], SAMEPERIODLASTYEAR ( Dates[Date] ) )

Sales YoY % = DIVIDE ( [Total Sales] - [Sales LY], [Sales LY] )

3Tableau: dashboards & storytelling

Tableau is the other industry-standard BI tool, loved for fast, drag-and-drop exploration and polished dashboards.

The core workflow

• Connect to data (a file, a database, or a published source).
• Build worksheets by dragging fields onto Rows, Columns and Marks.
• Calculated fields add new logic, e.g. Profit Ratio = SUM([Profit]) / SUM([Sales]).
• Parameters let viewers change inputs (e.g. a Top-N slider) interactively.
• Dashboards combine worksheets; Story points sequence them into a narrative.

4Automating Excel reports with Python

Many businesses still run on Excel. Instead of rebuilding a report by hand each week, generate it with Python — formatted, accurate and instant.

import pandas as pd

sales = pd.read_csv('sales.csv')
summary = sales.groupby('region', as_index=False)['amount'].sum()

# write a formatted, multi-sheet Excel workbook
with pd.ExcelWriter('weekly_report.xlsx', engine='xlsxwriter') as xl:
    sales.to_excel(xl, sheet_name='Raw', index=False)
    summary.to_excel(xl, sheet_name='Summary', index=False)

    book  = xl.book
    sheet = xl.sheets['Summary']
    money = book.add_format({'num_format': 'Rs #,##0', 'bold': True})
    sheet.set_column('B:B', 16, money)        # format the amount column

print('weekly_report.xlsx created')

weekly_report.xlsx created

5Scheduling automated reports

The final step in automation is making a report run by itself on a schedule, so it lands in stakeholders' inboxes without you lifting a finger.

# Linux / macOS — cron: run report.py every Monday at 8:00 AM
0 8 * * 1  /usr/bin/python3 /home/analyst/report.py

:: Windows — Task Scheduler (one-time setup, then it repeats)
schtasks /Create /TN WeeklyReport /TR "python C:\reports\report.py" /SC WEEKLY /D MON /ST 08:00

6Stakeholder communication & data privacy

Your final, highest-paid skill is communication — turning analysis into a decision a non-technical leader can act on.

Lead with the answer

Use the pyramid principle: state the recommendation first, then the key reasons, then the supporting detail. Executives want the “so what” in the first sentence, not on slide 14.

• Translate stats into business language: not “p = 0.01” but “we are confident this is a real ₹2L/month opportunity.”
• One message per slide; a clear title that states the takeaway.
• Anticipate the “why” and “what should we do” questions.

Handle data responsibly