	In Browser
	StumbleUpon
	del.icio.us
	Google
	Google Buzz
	reddit
	LinkedIn

	Facebook
	Twitter
	Linkedin
	E-Mail

Generative AI > Google Gemini API > Gemini in Google Sheets - AI Data Analysis Agent

Gemini in Google Sheets - AI Data Analysis Agent

Author: Venkata Sudhakar

Google Sheets is where much of Indian business data lives - sales reports, inventory trackers, expense sheets. Connecting Gemini to Sheets enables teams to get instant AI analysis without moving data into a separate tool. The agent reads sheet data, analyses it with Gemini, and writes insights back into a summary tab automatically.

In this tutorial, we build a ShopMax India monthly sales analyser that reads a sales data sheet via the Sheets API, sends the data to Gemini for analysis, and writes a structured insights summary back into a new sheet tab.

The below example shows reading sales data from Google Sheets, analysing with Gemini, and writing the summary back.

import os
from google import genai
from google.genai import types
from googleapiclient.discovery import build
from google.oauth2.service_account import Credentials

# Setup Google Sheets API client
SCOPES = ["https://www.googleapis.com/auth/spreadsheets"]
creds = Credentials.from_service_account_file("service_account.json", scopes=SCOPES)
sheets = build("sheets", "v4", credentials=creds).spreadsheets()

SPREADSHEET_ID = "your-spreadsheet-id-here"

def read_sheet_data(range_name: str) -> list:
    """Read data from a Google Sheet range."""
    result = sheets.values().get(
        spreadsheetId=SPREADSHEET_ID,
        range=range_name
    ).execute()
    return result.get("values", [])

def write_sheet_data(range_name: str, values: list) -> dict:
    """Write data to a Google Sheet range."""
    body = {"values": values}
    result = sheets.values().update(
        spreadsheetId=SPREADSHEET_ID,
        range=range_name,
        valueInputOption="USER_ENTERED",
        body=body
    ).execute()
    return {"updated_cells": result.get("updatedCells")}

# Read March 2025 sales data
# Sheet format: Date, Product, Category, Units Sold, Revenue (Rs), Region
raw_data = read_sheet_data("Sales!A1:F50")

# Convert to CSV string for Gemini
csv_data = "\n".join([",".join(row) for row in raw_data])
print(f"Read {len(raw_data)-1} rows from sheet")

Now send the data to Gemini for analysis and write insights back,

client = genai.Client(api_key=os.environ["GEMINI_API_KEY"])

analysis_prompt = f"""Analyse this ShopMax India monthly sales data and produce a structured report.

Data:
{csv_data}

Provide:
1. Top 3 products by revenue
2. Best performing region
3. Month-over-month trend (if available)
4. Product category breakdown
5. Three actionable recommendations for next month

Format as clear sections with bullet points. Keep it concise.
"""

response = client.models.generate_content(
    model="gemini-2.0-flash",
    contents=analysis_prompt
)

insights = response.text
print("Analysis complete. Writing back to sheet...")

# Write insights to a new tab "AI Insights"
# Split insights into rows for the sheet
insight_rows = [["ShopMax India - AI Sales Analysis - March 2025"]]
insight_rows.append(["Generated by Gemini AI"])
insight_rows.append([""])
for line in insights.split("\n"):
    insight_rows.append([line])

write_result = write_sheet_data("AI Insights!A1", insight_rows)
print(f"Written {write_result["updated_cells"]} cells to AI Insights tab")

It gives the following output,

Read 49 rows from sheet
Analysis complete. Writing back to sheet...
Written 67 cells to AI Insights tab

--- AI Insights Tab Content ---
ShopMax India - AI Sales Analysis - March 2025
Generated by Gemini AI

TOP 3 PRODUCTS BY REVENUE:
- Samsung 55 QLED TV: Rs 18.2 lakhs (23 units)
- Dell Inspiron 15 Laptop: Rs 15.5 lakhs (25 units)
- iPhone 15: Rs 12.8 lakhs (16 units)

BEST PERFORMING REGION:
- Mumbai: Rs 42.3 lakhs (34% of total revenue)

PRODUCT CATEGORY BREAKDOWN:
- Laptops: 38% | TVs: 29% | Mobiles: 21% | Accessories: 12%

RECOMMENDATIONS:
1. Increase TV inventory - sold out in Delhi and Hyderabad
2. Run a laptop promotion in Bangalore - underperforming vs Mumbai
3. Bundle accessories with mobile purchases - low attachment rate

Schedule this script with Cloud Scheduler or a cron job to run automatically at month-end. Add a chart generation step using the Sheets Charts API to visualise the top-products data. For larger datasets, use Gemini with a CSV file upload instead of pasting data into the prompt to stay within context limits.

Send your comments, suggestions or queries regarding this site to [email protected].