In partnership with
Let me describe a morning that happens in thousands of offices across India every single Monday.
You open your SQL client. Run 4 queries. Export the results to CSV. Open Excel. Paste the data in. Apply some formatting. Add a pivot table. Update the chart. Save the file. Email it to your manager with the subject line "Weekly Report — Feb 10."
Time spent: 90 minutes. Insights generated: Zero.
You just spent your first 90 minutes of the week as a human copy-paste machine. And you'll do the exact same thing next Monday. And the Monday after that.
This is not analysis. This is data entry with extra steps.
The best analysts I've worked with spend maybe 5 minutes on recurring reports—because they automated them months ago. The rest of their Monday is spent actually thinking about the data.
Let me show you how to get there, step by step.
Level 0: You (Right Now)
Manual everything. You run queries by hand. You paste results into spreadsheets. You format things manually. You email files as attachments.
There's no shame in being here—everyone starts here. But if you've been doing this for more than 3 months on the same report, it is a problem.
The Rule: If you have done the exact same data-pull task more than 3 times, it needs to be automated.
Level 1: Scheduled Queries + Email
This is the easiest win. Most database systems (SQL Server, PostgreSQL) or Cloud Warehouses (BigQuery, Snowflake) let you schedule queries to run automatically.
The idea is simple: Your query runs automatically at 7:00 AM every Monday. The results get emailed to your manager. You don't touch anything.
How it works in practice:
You write a query using relative dates (this is the secret). Instead of hardcoding
'2026-02-01', you useCURRENT_DATE - INTERVAL '7 days'orDATEADD(WEEK, -1, GETDATE()). That way, it always pulls "last week."You schedule it to run every Monday.
You configure it to send the results as an attachment.
-- The Weekly Sales Query (SQL Server Example)
SELECT
region,
COUNT(DISTINCT order_id) AS orders,
SUM(revenue) AS total_revenue,
ROUND(SUM(revenue) / COUNT(DISTINCT order_id), 2) AS aov
FROM orders
-- The magic is here: No hardcoded dates
WHERE order_date >= DATEADD(WEEK, -1, CAST(GETDATE() AS DATE))
GROUP BY region
ORDER BY total_revenue DESC;
Reality Check: As a junior, you might not have the database admin rights to schedule jobs directly. If so, simply hand your dynamic query to the Data Engineering team/DBA and ask them to schedule it, or use the "Email Subscription" feature built into tools like Metabase or Redash.
Time saved per week: 60–90 minutes.
Level 2: Dashboards with Auto-Refresh
Level 1 gives your manager a static email. Level 2 gives them a live dashboard they can check anytime. This is where Power BI, Tableau, or Looker earn their keep.
The setup:
Connect your BI tool directly to the database (not to an Excel file—that defeats the purpose).
Build the dashboard once with all the charts and filters your manager needs.
Set up a Scheduled Refresh (e.g., Power BI Service lets you refresh up to 8 times a day on Pro licenses).
Now, your manager doesn't wait for your Monday email. The data is always current.
The mindset shift: You're no longer a report-runner. You're a dashboard-builder. Your manager self-serves the basics. They only come to you when they need something the dashboard can't answer—which is when the actual, interesting analytical work begins.
(Common junior mistake: Building the dashboard but still manually clicking "Refresh Data" every morning. Configure the cloud refresh!)
Time saved per week: 2–4 hours (because it also kills all those ad-hoc "can you pull this number for me?" requests).
A quick shoutout to our sponsor who keeps this newsletter free. It helps us keep the lights on if you check them out. (Continue Reading). STOP TYPING and reduce strain on fingers, try it for free
Better prompts. Better AI output.
AI gets smarter when your input is complete. Wispr Flow helps you think out loud and capture full context by voice, then turns that speech into a clean, structured prompt you can paste into ChatGPT, Claude, or any assistant. No more chopping up thoughts into typed paragraphs. Preserve constraints, examples, edge cases, and tone by speaking them once. The result is faster iteration, more precise outputs, and less time re-prompting. Try Wispr Flow for AI or see a 30-second demo.
Level 3: Python Scripts That Do Everything
This is where things get powerful (and where people tend to overcomplicate). You don't need a data engineering degree. You just need one Python script that does 3 things:
Connects to the database and runs your query.
Does complex transformation (cleaning, logic, formatting).
Sends the output somewhere (Email, Slack, Google Sheets).
Here's a real example. Every Wednesday, your marketing team wants a "Reactivation List" of customers who placed their first order last week but haven't returned.
import pandas as pd
from sqlalchemy import create_engine
# (smtplib / email imports omitted for brevity)
# 1. Pull data directly from Postgres
engine = create_engine('postgresql://user:pass@host/db')
query = """
SELECT customer_id, email, MIN(order_date) AS first_order_date
FROM orders
GROUP BY customer_id, email
HAVING COUNT(*) = 1
AND MIN(order_date) >= CURRENT_DATE - INTERVAL '14 days'
"""
df = pd.read_sql(query, engine)
# 2. Transform (Calculate days since order)
df['days_since'] = (pd.Timestamp.now() - pd.to_datetime(df['first_order_date'])).dt.days
# 3. Export and trigger email
df.to_csv('/tmp/reactivation_list.csv', index=False)
# Code to email the marketing team goes here...
You schedule this script using a cron job (Linux/Mac) or Task Scheduler (Windows). Every Wednesday at 8 AM, marketing gets a fresh CSV. You never think about it again.
Time saved: Whatever it took you to do this manually... forever.
How to Pitch Automation to Your Manager
Don't say: "I want to learn Python automation." Say: "I currently spend 6 hours a week on recurring reports. If I take 2 days to automate them, I'll free up 6 hours every week for deeper analysis on our drop-off rates. That's 24 hours a month back."
Managers don't care about your technical growth. They care about business output. Frame it as output.
The Trap: Automating the Wrong Things
Not everything should be automated.
If a report is requested once and never again, don't build a pipeline for it.
If the business logic changes every month, a rigid Python script will break constantly and create more work.
Automate the stable, repetitive, boring stuff. Keep the exploratory, ambiguous stuff manual. That's where your actual analytical brain adds value.
One Thing to Do This Week
List every recurring task you do (weekly reports, daily pulls). Write down how long each takes. Pick the one that is the most tedious and least intellectually stimulating.
That is your automation candidate. Start with Level 1. Get the win first.
Next week: Price Sensitivity 101: the analysis your manager will ask for, and you won't find on Kaggle.
— Ani Real Analyst Jobs
P.S. Added new 35+ fresh remote jobs: realanalystjobs.com
Added new free projects : https://realanalystjobs.com/projects
Document your 2026 Journey: https://realanalystjobs.com/journey
Talk to me : https://realanalystjobs.com/raj