1. Export Your Google Fit Data#

Head to Google Takeout and export your Fit data. You’ll get a zip file containing daily CSV files — one per day — with columns for steps, heart rate, calories, distance, oxygen stats, and more.

Each file looks something like this:

Start time,End time,Move Minutes count,Calories (kcal),Distance (m),
Heart Points,Heart Minutes,Avg heart rate,Max heart rate,Min heart rate,
...,Step count,...

The column we care about is Step count — but it’s buried among 20+ other fields. That’s why we need the next step.

2. Clean the Data with a Jupyter Notebook#

I used Google Colab to write a short notebook that reads all 31 daily CSVs and extracts just the step count per day.

Cell 1 — Import pandas:

import pandas as pd

Cell 2 — Loop through all 31 days and aggregate:

final_df = pd.DataFrame(columns=['date', 'steps'])
for i in range(31):
    date = f'2026-08-{str(i+1).zfill(2)}'
    df = pd.read_csv(f'/content/{date}.csv')
    total_steps_of_the_day = int(df['Step count'].sum())
    new_row = pd.DataFrame({'date': date, 'steps': total_steps_of_the_day}, index=[0])
    final_df = pd.concat([final_df, new_row], ignore_index=True)

Cell 3 — Preview the result:

final_df

This outputs a clean table:

Cell 4 — Export to CSV:

final_df.to_csv('/content/Steps_08_26.csv')

The output is a tidy two-column file: date and steps. Total for August: 472,434 steps.

3. Upload to Manus AI and Give It a Prompt#

Open Manus.im, upload Steps_08_26.csv, and tell it what you want. Here’s the exact prompt I used:

I want Manus to fill up the fundraising details on a website for Children with Cancer UK with my steps details. Here’s the access link [URL]. You have the step details in the CSV.

That’s it. One paragraph. Manus takes it from here.

4. Answer Manus’s Clarifying Questions#

Manus didn’t blindly start entering data. It read the CSV, inspected the fundraiser site, and came back with two smart questions:

Question 1 — Unit mismatch:

The activity form accepts miles, not steps. Your CSV records 472,434 steps from 1–31 August 2026, but it does not contain a distance or a steps-to-mile conversion. Please tell me the conversion you want used.

My answer: 2,000 steps = 1 mile

Question 2 — Entry format:

Option A — One cumulative entry of 236.22 miles dated 31 August. Option B — 31 separate entries, each dated and converted from that day’s steps.

My answer: B (daily records)

This is exactly what you want from an AI agent — it identifies problems before they become mistakes and gives you clear options.

5. Let Manus Do the Work#

After my answers, Manus:

1. Wrote a Python conversion script using Decimal with half-up rounding to avoid floating-point errors:

from decimal import Decimal, ROUND_HALF_UP

STEPS_PER_MILE = Decimal('2000')

for record in csv.DictReader(source_file):
    exact_miles = Decimal(steps) / STEPS_PER_MILE
    displayed_miles = exact_miles.quantize(Decimal('0.01'), rounding=ROUND_HALF_UP)

1. Generated a daily mileage schedule — a markdown table with date, steps, and miles for all 31 days:

1. Posted all 31 entries to the Children with Cancer UK fundraiser site — one by one, each with the correct date and mileage. Every submission returned HTTP 200.

6. Verify the Result#

After all 31 entries were posted, the live Activity challenge tracker displayed 236.23 miles — matching the expected sum of the 31 individually rounded daily values.

• ✅ 31 entries posted
• ✅ All HTTP 200 responses
• ✅ Public tracker shows correct total
• ✅ Login token expired after completion (task done)